 Hello, everyone, and welcome to this joint meeting of the Geographical Sciences Committee and the Mapping Science Committee of the National Academies of Science, Engineering, and Medicine. I am Harvey Miller, and I'm Chair of the Mapping Science Committee. We're glad to have you with us whether in real time on Zoom or viewing the recording of this session later. Now, before we begin today's webinar, I'd like to say a few words about the National Academies of Science, Engineering, and Medicine and the Mapping Science Committee. Next slide, please, on the expertise of volunteers to inform policy with evidence, spark progress and innovation, and to confront challenging issues for the benefit of society. Next slide, please. The Mapping Science Committee, a standing committee within the National Academies, was established in 1987. It addresses aspects of geographic information science that deal with the acquisition, integration, storage, distribution, and application of geospatial data. The MSC tracks and hosts discussions on mapping and geospatial industry initiatives and advancements with the goal of increasing collaboration and synergy among business, government, and academia. The committee also tracks international mapping and geospatial science research and technologies that may have value to the nation. The current members of the committee are Stuart Fatheringham, Oceana Francis, Hendrick Hammond, Kristen Kerlin, Marguerite Madden, Keith Mossback, Kathleen Stewart, and Eric Tate, and we greatly appreciate their time and energy in helping the MSC be successful. Our sponsors are the U.S. Geological Survey and the U.S. Census, and we are grateful for their support. I'd now like to pass the mic to the chair of the Geographical Sciences Committee, Dr. Pat McDowell. Please slide, please. Thank you, Harvey. The Geographical Sciences Committee provides advice to society and to government at all levels using the methods of spatial analysis and representation. We address the geographic dimensions of human-environment interactions, spatial location and concentration, and place-based research and policy at all spatial scales. The committee also fosters international cooperation by serving as a liaison to other national geographical organizations, including as official U.S. liaison to the International Geographical Union. The current members of the committee are Boudindra Baduri, Janet Franklin, Janelle Knox-Hays, Glenn McDonald, Elizabeth Root, and Dawn Wright. And thank you all so much for your service. The committee is supported by funding from the National Science Foundation. Next slide, please. In our meeting today, we will discuss ways to integrate qualitative or thick data with quantitative data to construct a real-time understanding of forced migrations. We will also explore the gaps that need to be filled to improve this understanding, as well as the limits, opportunities, and errors associated with both qualitative and quantitative geographic data. We'll examine what qualitative data is available and the ethical issues around using these data. Finally, we'll discuss how we can visualize and communicate data sets that integrate qualitative and quantitative data. Our program today will consist of four sessions, a keynote, and three panels. The committee has prepared a few questions for each session, and will also take questions from the audience. So please feel free to enter your questions into the Q&A box at any time. Please note that this webinar is being recorded, and understand that any questions you submit may be read aloud and included in our recording. A link to the recording will be posted on the MSC and GSC websites within the next few days. Now, I'd like to introduce our keynote speaker, Elizabeth Fussell. Dr. Fussell is Professor of Population Studies and Environment and Society at Brown University. She's a sociologist and demographer. Her research focuses on environmental drivers of migration and social inequalities in migration, health, and other post-disaster outcomes. Her talk today will focus on climate migration, displacement, and relocation. Dr. Fussell. Thank you. Thanks for inviting me here today. So I need to share my screen. There we go. Great. So I join you today from Providence, Rhode Island. And this is on lands which are within the ancestral homes of the Narragansett Indian tribe. The university acknowledges that the Narragansett Indian tribe was dispossessed from their lands by the forces of settler colonialism, and we acknowledge our ongoing responsibility to understand and respond to the legacy of those actions. This land acknowledgement is a really good starting point for this keynote address because it reminds us of the long-lasting harm of forced displacement. I'm going to begin today's talk with my own story about displacement that launched my research on this subject. I was an assistant professor in the sociology department at Tulane University in New Orleans when Hurricane Katrina made landfall on August 29, 2005. The day before, as faculty, staff, and students were preparing for the new semester, a mandatory evacuation was called for the entire city of New Orleans and the lower-lying parishes. My husband and I gathered our one-year-old daughter and our two cats, and we headed to Baton Rouge. From Baton Rouge, we traveled to Northern Virginia, Philadelphia, Pennsylvania, and then back to New Orleans in time for the Christmas holiday. While on this journey, I wrote an essay for the Social Sciences Research Council's website, which was titled, the website was titled, Understanding Katrina. The website gathered essays from social scientists who study disasters, environment, and society. My essay, which is shown here on the slide, focused on social stratification, social networks, and the hurricane evacuation. All of these are topics which I continue to research. While I was in Philadelphia, the SSRC convened a meeting in New York City of several of the essay contributors, as well as interested members of the public. As a migration scholar, I was new to disaster research, and I listened quietly and respectfully as scholars who had built their careers studying disasters debated whether it made sense to rebuild New Orleans. I immediately imagined all the reasons why rebuilding New Orleans made lots of sense. It is not uncommon to hear such comments about risky places after an especially devastating disaster. But this view overlooks all of the personal and societal loss that residents experience. And fortunately, one of my displaced friends who was in the audience, he's a Louisiana native who worked as a culture writer. He reacted quickly to the suggestion that New Orleans shouldn't be rebuilt. He spoke passionately about the meaning of New Orleans and other parts of the devastated Gulf Coast and all of the millions of residents who built their lives there. As a resident and a researcher, I understood both perspectives. And we might think of each perspective as privileging different types of data. Researchers have a multidimensional and historic relationship to a place that is not easily summarized in any metric. We can characterize that information from residents about their identities and their lived experiences as qualitative. Researchers and contrasts seek to reduce information into a set of measures that statistically explain a large proportion of the variation in a range of outcomes. This is an inherently quantitative approach that privileges concepts that can be quantified such as property loss and economic costs and neglects the qualitative concepts that are more difficult to measure like culture and community. Those of us attending this workshop today recognize the societal benefits to be obtained from integrating qualitative and quantitative data to produce timely information on forced migration. Such information is key to developing responsive disaster recovery policies that produce an equitable outcome for all of the disaster affected residents. So today I have two objectives for my talk. First I will orient you to conceptual issues in measuring forced migration. And these are the sorts of conceptual issues that preoccupy migration scholars. And I'm going to review data sets that have been used to estimate post-disaster migration quantitatively and then talk about some of the challenges for combining it with qualitative data. I hope that this somewhat inconclusive ending will launch a good discussion about how to move forward. So turning now to concepts used in migration research, I'm focusing on this figure that comes from the United Kingdom Government Office for Sciences foresight project report. The foresight project sought to understand how global forces, but especially climate change, will affect the volume and patterns of human migration in the near future and how policy makers might address the root causes of migration. It's kind of a complex figure, I'm not going to go over the whole thing, but I'm just going to summarize a few of the widely agreed upon findings from the report on environmental migration. And we can define environmental migration as any migration influenced by environmental changes and events. So one of the first conclusions is that environmental changes and events, whether those are temperature and precipitation extremes, sea level rise or extreme weather events, rarely have a direct effect on migration. They typically operate through the other drivers, which are represented here in the left hand side of the panel, as the five points on the Pentagon. And the especially relevant are the economic drivers, the environmental drivers and the political drivers. Second, most migration influenced by the environment occurs within countries and usually over very short distances. This is true of most migration in general, regardless of the motive for moving. Violence and political persecution are exceptional because those are the main reasons for international migration to flee political persecution. Finally, environmental changes and events are as likely to prevent as they are to produce migration. We see this on the right hand side of the panel, which shows that contextual variables and household and individual level characteristics influence the migration decision. Therefore, migration is selective, meaning that certain types of people are more likely to move and others are more likely to stay. These findings can be reduced to a stylized fact about environmental migration. It is rare, usually short distance and selective on individual and household traits. But what this figure doesn't show very well is that the migration decision is not always a completely free choice. Migration scholars are more attentive to how structure and agency interact in the migration decision. We define migration as any residential move by an individual or household that is permanent and crosses a geographic boundary. This broad definition of migration is inclusive of many subtypes of migration, which can be classified along three dimensions represented by this cube, distance, duration, and volition. I've shaded the back of the cube to represent migration at the involuntary extreme of the volition dimension, and the x-axis represents duration and y represents distance. Each dimension is a fairly intuitive concept, but the challenge is to operationalize them. For example, the distance between origin and destination might be measured as which political boundaries are crossed or linear distance or travel time. Similarly, duration might be measured by the time the migrant is observed to have lived in the new destination, in other words, how long they've lived in a new destination, or by the migrant's intention for staying in that place. Volition is rarely measured directly because it requires the migrant to state their reason for moving. Usually, volition is inferred from the circumstances under which a move occurred. So most residential moves are permanent, short-distance, and voluntary. These are moves that happen alongside life course changes like marriage, cohabitation, separation, divorce, widowhood, and the search for housing appropriate to the changing size and needs of the household. Thus, they're usually very local. Longer-distance permanent migrations are more often associated with employment and may fall somewhere in the middle of the volition dimension, depending on the options for in terms of employment. For example, college professors are recruited to work at a university, but not necessarily the university in the city which they prefer to live in. And so that's a somewhat constrained choice or migration choice. So our interest is in migration that is forced, that is driven more by the push factors than the pull factors. And these might include hazard evacuation, which would be an involuntary move that may not even be a migration because it lasts only a short period of time and involves only a short-distance move. But an evacuation may become a displacement when destruction of the built environment is extensive and it takes a long time to rebuild the pre-disaster home or to find new housing. A displacement may become a permanent relocation when land loss or hazard exposure prohibits return to a pre-disaster home or community. Long-distance moves across an international boundary are most likely when people are fleeing political persecution or conflict and may range in duration from temporary to permanent. The environmental drivers rarely contribute to such migrations. These types of migration often have push factors that can be temporarily defined, making it easier to infer the reason for moving from those circumstances. And therefore it's easy to label these as forced even without talking to a migrant who would provide you with the reason for moving. However many migrations occur in the ambiguous interior of this three-dimensional box where it is difficult to isolate a single driver of migration and that was the point of the of the foresight slide that showed that most drivers operate or the environment tends to operate through other drivers of migration. So these maybe moves in response to gradual changes in the environment or the context like a drought, repeated environmental hazards, diminished natural resource-based livelihoods, or political repression. In these gradually worsening conditions, potential migrants search for better employment or educational opportunities or something else, better housing. They search for that elsewhere and thus they combine the push and the pull factors in them in their migration decision. All of these spaces within the three-dimensional conceptualization of migration are types of forced migration and each involves some degree of choice about whether, when, and where to move. So let's return to the conceptual framework. What I've just discussed about the temporal, spatial, and volitional dimensions of migration shows that the migration decision is not just whether a person decides to migrate or to stay in place. The decision also involves when to leave, where to go, whether to return, whether to settle in the new destination, and whether to continue migrating. All of these decisions, according to this model, depend on the the macro, the mezzo, and the micro level factors that are identified in this figure. These factors are the structures and forces that constrain the migration decision by opening up some migratory pathways and closing off others. They allow potential migrants to exercise some degree of agency even if there are very compelling push factors. So let's think about this model relative to the example of Hurricane Maria's effect on Puerto Rico. The hurricane destroyed housing and caused prolonged power outages throughout the island, causing many residents to consider whether they would be better off elsewhere, at least temporarily. Puerto Rican status as U.S. citizens provided a pathway to the states and social networks connected some of them to friends and family in specific places like Miami, Orlando, New York City, and other cities where Puerto Ricans had previously settled. Data from multiple sources estimate the population of Puerto Rico to have declined by about 100,000 residents between 2017 and 2018. So we infer that about 100,000 people moved permanently from Puerto Rico to the U.S. states. This is a dramatic loss for sure, but it's only around 3% of the island's population of 3.3 million. Most Puerto Ricans did not migrate. They stayed and rebuilt their lives and communities. Outmigration was greater from the more devastated areas, and some quantitative evidence shows that migrants were more willing to be working, sorry not more willing, more likely to be working age adults who were most likely seeking work in the U.S. to support the family that they had left behind. So what this case illustrates is that a humanitarian response to a crisis that has the potential to cause migration would be a policy that preserves and enhances residents' agency and allows them to exercise control over their residential outcome. In this case, migration was one of a set of possible responses to the crisis in Puerto Rico, but clearly not the only response. The well-traveled pathways to familiar destinations provided an option for many Puerto Ricans who were struggling with prolonged disaster conditions on the island. And we can imagine other examples where the migration option might have been more or less attractive depending on the push factors, but it's important to remember that even in a situation where we that we label as a forced migration, potential migrants or the actual the people who do actually migrate have a lot of agency and that agency needs to be recognized and protected. So staying with the example from Puerto Rico, I'm going to wrap up with a review of the data sources that were used to study migration from Puerto Rico after Hurricane Maria. I assembled this table listing each source of data as well as five characteristics on which they can be compared. The major differences between the first four types of data listed above the black line and the last two types of data listed below. The first four types of data come from administrative records that record movement at specific time intervals. For example, geotagged tweets from Twitter measure current locations whenever an individual tweets. So you can tell if a person's moved from one location to another. IRS records also record residential addresses at the time of filing and we can tell by comparing addresses from one year to the next whether a tax filer has moved or remained at the same address. These changes in locations are what we count as migration and they obviously occur at different time intervals. But these data are not representative of an entire population of a geographic area. They're selected by who is tweets, by who files taxes. So that's one of the limitations. They also have few or no demographic measures. We don't know anything about tax filers from the data that's released in this IRS migration flows dataset except how many people, how many dependents are in the household and what their adjusted births income is. However, this type of data from administrative records is available relatively quickly and can often be accessed for free or with low cost. The last two lines of data describe data from federal or specialized surveys. These are time intensive to collect and they cost lots of money. They produce cross-sectional data for the American Community Survey. They're reported annually for a specialized survey that somebody might, like a researcher, might carry out. They might only occur once or maybe twice but it's hard to collect longitudinal data in that way. These survey data are rich in qualitative measures of individual and household characteristics and they, if done correctly, are representative of the population. They don't typically include narrative data about lived experiences. So that's a different level of qualitative data that is would be even more difficult to collect at this scale. In these surveys, migration is measured retrospectively with a question about where a person lives currently and where they lived at some point in the past. Surveys are slow so they're not going to be useful for tracking migration in real time but they have the potential to tell a much richer story about who migrated, where they went, and the relative well-being of migrants compared to non-migrants. Survey data can be combined with remote sense data in a statistical framework to describe in some times even do a causal analysis of migration. So, you know, these are rich data that we use to do a lot of research but they're not going to be timely. So the takeaway from this exercise is that there's a trade-off between the timeliness of data and the availability of qualitative measures. I'm going to end here with one final thought to focus the work we're going to do today. This thought harkens back to the lesson that I learned from my experience as a Hurricane Katrina survivor and a scholar as well as the extensive case study literature on forced migration. In disaster situations residents have agency and want to make choices even when migration seems to be the only option. A humanitarian response to forced migration is one that expands and enhances migration's options of when and where to go, how long to stay, and with whom they will migrate. I'll end things there. I look forward to your questions. Thank you very much. Thank you, Elizabeth, for a really interesting talk that really gave us a lot of structure to start this discussion off. I appreciate it very much. I'm going to ask the committee members and other panel members if they have any questions at this point. Is there anybody that has a question for Elizabeth at this point? I'd like to ask a question if I could, Pat. I'm interested in the last line you said about humanitarian responses to open options, opening options for migration and displacement. I was wondering if you could describe the type of data that would be useful for that. Well, I think that the, I've said that's the type, I don't think there's a single type of data that's useful for supporting that kind of assertion. You know, I referred to a broad literature on with case study research on lived experiences of forced migration. I think that when you look at say people who have had forced migration experiences where they end up in an internally displaced person's camp where they have very limited options and very little control over the circumstances of their lives, these are some of the most miserable places on earth. I mean they're just, and you know that literature is, you know, it's out there. It's out there that these camps are very undesirable. In the United States our disaster assistance programs are designed to enhance people's options for residential, post-disaster residential mobility. And I've done, I've researched that kind of, that data on that myself. We do, I have, I'm part of a project called the Resilience and Survivors of Katrina Study that where we did in-depth interviews as well as survey research with a cohort of women living in New Orleans before Hurricane Katrina. And one of the things that we find is that there's, the people who had longer displacement trajectories had disrupted lives for a longer period of time whereas those people who were able to go home relatively quickly were able to recover and restore themselves to that pre-disaster condition much more quickly and that was just simply better for people. So there's a lot of data that can bear on this question of what's the, what's the best kind of post-disaster migration outcome. And it wouldn't be any single type of data. But you seem to be emphasizing quality of data interviews, things that get out this lived experience and these life courses that led to this, this response to the disruption. Yeah, for sure, for sure. Okay, thank you. Great, thanks Harvey. Any other questions from the panel at this point? One of the audience members asks a question about spatial scale of analysis and the question is what higher level units of analysis might be used in migration studies? For climate change relocation, for example, might neighborhood level data or city level data be used? And would there be, you know, different kinds of insights and advantages to aggregating up to that level? Yeah, I've used the IRS county to county migration flow data set to look at post-Katrina residential mobility and we have been able to show that, you know, we've done a lot of research on what we're, what me and my colleagues call recovery migration that shows kind of the stage, the process through which people left New Orleans and then moved back to New Orleans in the subsequent years. And the reason why we can use county level data for that, for that study is because the Hurricane Katrina's destruction was so massive. And so it happened, you know, the entire parish of New Orleans, not the entire, but there was a mandatory evacuation for the entire parish and 80% of the city was flooded. And so the displacement scale just for that single place was massive. And so, and it's also, there were a lot of people living there. So it was easy to use this data to trace migration patterns. But most disasters have a smaller footprint than Hurricane Katrina. They would occur at the sub-county scale and they would affect fewer people. And so that approach, that spatial scale probably wouldn't work for something like a tornado. So then you'd have to ideally have tracked level data, census tract level data, or some other kind of spatial scale that would be smaller and more appropriate to the footprint of the hazard or the disaster. Yeah, that makes a lot of sense. I imagine there is some gap in the depth and the maturity of research on large-scale disasters versus small spatial-scale disasters. Is that correct? Are these sort of small dispersed individual point disasters less understood and less known? Yeah, for sure. I mean, we have a total, there's a very big bias in social science research on disasters toward the big ones. You know, those are the ones that make the headlines that capture our attention and that produce enough sort of impact or statistical power that we can do a survey and actually expect to have some statistically significant results. And we don't know very much about the impact of smaller disasters. My colleagues and I have also used county level population change data to look at the entire range of hurricane events in the United States and their impact on county level population change. And one of the things that we found that I think is super important to recognize is that most disasters or most hurricanes, let me, sorry, they're different. Not all hurricanes produce a disaster, but most hurricanes actually do not register any effect on the change and on the population size of the county. In fact, most, when a hurricane does affect the size of the population of a county, the effect tends to be to depress growth in counties that are growing. But most counties in the U.S. are rural counties that have stable or declining population sizes. And in those counties, they typically don't register a statistically significant effect. So it's really just the large growing counties that are most affected by a hurricane impact. And actually sometimes the effect is a little counterintuitive because after a hurricane there's a huge infusion of recovery money. And sometimes that actually attracts people to a place so that in some instances there's actually population growth after a hurricane, particularly if that place was already on an upward growth trajectory before the hurricane. Do you have any thoughts on what should be done to address this gap in understanding between large disasters and spatially smaller disasters? Do you have suggestions for kinds of research that should be started or new datasets or techniques that should be used? Yeah, I do. This is one of the concerns that my colleagues, Jack DeWard and Catherine Curtis and I share. We have a research agenda around this where we're trying to assemble data on the entire population of, say, hurricanes or tornadoes or floods and then map them onto population with a spatially temporally harmonized longitudinal dataset to understand how these statistical relationships look. That's a research agenda that we have underway. Unfortunately, we are constrained to the county scale because of the way population data and disaster data are organized. So I think one innovation would be to figure out how to combine that data at a smaller spatial scale. I think Census Bureau would only release it at the Census tract level and so that's, you know, we're somewhat constrained by that. But understanding the sort of generalizable effects in the U.S. of hazard events at whatever scale and their effect on population change is one that's definitely where we are right now, where we need to keep pushing. The other thing I would add is that one of our findings that we haven't yet published from this research is that places are adapted to the hazards that they usually experience. So places that experience more hurricanes actually experience less population change after a hurricane. And places that experience, you know, no hurricanes, obviously they don't cause population change. It's that middle range of places that infrequently experience a hurricane and therefore may not be well adapted to that hurricane. They're going to experience more damage and more of the kind of displacement that happens when a hurricane is particularly destructive. So that's another feature of places that needs to get incorporated into this sort of study. Places are not all equal in their ability to cope with these natural hazards. Yeah and I suppose that idea could be translated to different types of disasters. So for example this event that we're hearing about in Yellowstone National Park in the news today, maybe they'll label it a 100-year flood. I'm not sure if that's what they'll label it, but it seems to be a very unusual event for that location. And the big floods that really affect people's lives usually are quite rare. So there may be classes of disasters that are at one end of this scale and other kinds of disasters that are at the other end. That's really interesting. Yeah I mean the way I think about it is that in the United States we have for over a hundred years the Army Corps of Engineers and localities have been hardening and managing the environment, hardening our infrastructure to mitigate hazard impacts. And those those investments in protection have, they work, they really do work. And what we're experiencing right now with climate change is an increase in the number of extremely destructive events for which that infrastructure, those investments may not be prepared. It may be hitting the maximum capacity to withstand those types of increasingly frequent extreme events. And so we either need to continue adapting and improving that infrastructure or we need to decide where we're going to move people out of harm's way. And so this sort of, there's a national discussion about managed retreat and how that's going to unfold. I think that my point about making sure that policies like managed retreat respect residents agency and give them tools to make the to do managed retreat on their own terms as opposed to a top-down approach from the government is really important for avoiding a humanitarian crisis. So that would be, that's where I think our conversation is right now where a lot of people are increasingly using this word managed retreat to think about how we will cope with the increasingly hazardous future. And so kudos to you all for being forward thinking about how to make sure that that we avoid these sort of human rights violations that honestly we've seen in the United States in the past or even right now when you think about places like Il-de-Jean-Charles in Louisiana. That has been an important object lesson in how to I'm not going to say it's been a hundred percent successful, but there are a lot of lessons. Let me just put it this way. There are a lot of lessons to be learned from that particular case. Great. There's a couple of other questions. I'm not sure we'll get to all of them, but here's another one related to the data issue. Has there been any use of information from citizen science, community-based science, do-it-yourself efforts or contacts made through organizations such as co-op extension services, etc. that have been useful in understanding these issues? I can think of a few of those sorts of you know citizen science or do-it-yourself research projects that have been used to study disaster effects on communities. One of those effects being displacement or evacuation, they are I'm sorry I can point you to that study later. I'm not thinking the name of the author right now, but there was one that was done after Hurricane Katrina. There are actually several that were done after Hurricane Katrina. I think the issue is that again that sort of data collection effort happens for the big events and that when a small event occurs we may not see the same kind of mobilization of resources to collect this kind of information. Yeah great. Thank you. Well thank you Elizabeth so much for that excellent opening discussion for the day. Now I'm going to hand the mic over to Mapping Sciences Committee member Kristen Kurland to moderate the first panel. Thank you very much Pat and again thank you Elizabeth that was an excellent talk and I learned a lot. In our first session we will hear from three speakers about using human-centric geographic data to map human movement and land use and also design humanitarian response efforts. Each speaker is going to speak for about 15 minutes and we'll hold questions from the committees and those of you listening throughout the webcast until we've heard from all the panelists. So I'd like to welcome our first speaker who is Anna Triandafiliadu. She's a sociologist and mitigation policy expert at Toronto Metropolitan University so take it away Anna. Hello. Hello everyone. Oh yeah I was about to say okay you have my powerpoint right. Okay thank you all and thank you Elizabeth for everything. I mean I have a few questions but I thought perhaps some of them will be answered today as we as we all make our presentations. So well delighted to join you I'm joining from Toronto that is the dish with the one spoon territory the traditional land of the Anasinabe, the Haudenosaunee and the Mississauga tribes and this with one spoon is actually also very pertinent to our to our discussion today because the dish is our natural landscape our natural resources and we all have one spoon to sharing them both synchronically like the people who we are here today and we're all invited in peace to share but also I feel that this is a diachronic message in terms of us having taken those resources from previous generations as Tommy Memorial and the Indigenous peoples and having the responsibility to transmit them to the next generations and to keep them in good order. And having said this I just want to then go back to my the main how can I say focus of my talk which is how do we use data and actually how do we deal with the very extensive availability of data and that's where I mean I like very much some of the things that Elizabeth said and as I said I have questions also in terms of who provides the data who uses them you know and how do you know how do we incorporate them in our analysis if we can go to the next slide because I believe you have the control right I don't yeah so actually I want to start by just showing a few slides about how we use migration data and what kind of data we have available and alert us to who creates those data so this is an example of the data that we had from the so-called Balkan path in during the refugee emergency in the Mediterranean and in southeastern Europe and the whole of Europe in 2015-16 and I want to just add the personal reflection as a researcher at the time I was so delighted to have real time data so you could go on that website either of IUM or the UNHCR and you could see every week how many men women children what nationalities were crossing mainly through Greece but also through Italy or Spain every day and you could even sometimes actually trace where they were going there was no effective you know tracing in terms of following people but you can see those graphs that were being produced and I was also producing and reproducing them and using them in my own research if we can go to the next slide that's part of the same you know actually an outgrowth of that initial coverage of the 2015-16 you know what I I personally call it the refugee emergency generally has been known as the migration crisis and that's also something to to be debated who's crisis what crisis but anyhow I think that part of that and part of the concern with people losing their lives in crossing the Mediterranean give birth to the missing migrants project which is an IUM project and here for instance you can see I mean this the screenshot is from a current how can I say database being constructed in relation to displacement in Afghanistan and I chose specifically this snapshot rather than the map that you will see later on on the Ukraine current you know displacement and refugee crisis because one would think you know on one hand you're thinking oh that is so wonderful we know what is happening but then you have to ask yourself how do we know that how can we know how many people were intercepted how many people arrived in Europe how many attempted crossings how many deaths and disappearances so I think that is a very important point that like who is collecting this data whose data are those what do they show and how do we use them and if we can go to the next slide sorry actually I have sorry I have to to to correct myself the previous slide was still on the Mediterranean crisis I'm going I'm going to go back to the to the Afghanistan thing in a moment so so but but my questions are still valid in terms of who collects them and how do we know so and here you can also see obviously this this big capacity of collecting data that we have is reflected currently on on the Ukraine and and you can see those data that you can access in real time and actually with this really nice visualization of you know where are the so obviously the big bubbles are the many displaced and refugee people and and you can also have this different accounts so for instance refugees but also border crossing so at least there is now an awareness that it can be more border crossing than people because people can can cross more than once or different borders and if we can go to the next slide so and then here is the the the Afghanistan case and I think I'll I'll want to to spend a minute on the slide because this is the IOM displacement matrix so basically IOM has local people from their IAM branches in different border crossing points and they collect the data of the people who are crossing now we have with a with a colleague of mine Yuna Sahuga who is studying more IOM and the global governance structures so he was alerted me to the fact on how you know on how these data are collected and there is apparently 200 people in Geneva who are processing this data but we are still in the process and I don't want to say that we're going to find something how can I say irregular or unethical but we're still in the process of digging to find out because my question to my colleague was immediately under whose jurisdiction are they collecting this data do they have ethics and I'm not someone you know that this big ethics bureaucrat but suddenly I was thinking this data are collected from people in distress because they're crossing a border because they're fleeing so these are recent data on on Afghanistan after after august 2021 so we're talking about borders that it's not like you know the I don't know the the Spain the Spain-France border at the Pyrenees or not the the German German Polish border not the US-Canada border so these are borders that are probably in many places how can I say they're border areas rather than a very you know bureaucratic border of the kind we are imagining when we're crossing here from Canada to the US by car as I said people are in distress I am people who are collecting this data are probably to the best of my knowledge are collecting this data while they are also providing some information or whether about support whether about programs for repatriation or from further moving onwards and my question is the extent to which people are given an opportunity to say no I don't want to give my data I don't want you to register me whether the support they're provided is conditioned upon this data and actually this apparently this very systematic this first systematic collection that is now expanding in different world regions was something that was started after the Libya crisis nearly 10 years ago when actually IOM did certainly a very crucial thing because it was providing for support at the at the southern border of Libya for people from sub-Saharan African countries that were fleeing not only generalized violence and disorder in Libya but also specifically racist violence against them and there was an effort to try and help people to reach a safe haven and also repatriate to their the respective countries of origin so what I think I want to say is that data can be collected for positive reasons or for good reasons and we know that's actually a big motivation behind collecting such data has been also when people are going towards refugee camps to provide information you know there is this population arriving this population has children has babies you need to send milk you need to send diapers you need to send x amount of food but the problem is that today with the technology that we have this information travels very fast and can travel very far and it is clear that often the people that provide them are people in this stress and people who are not even if they were told you know you have the the the possibility to refuse to give your data maybe they actually don't have the responsibility the possibility and if we can go to the next slide so this is my my my these are my question marks in terms of real data and I want to again emphasize that particularly when in 2015-16 at least that was the first time I was aware that there was this possibility of having real-time data and I've written a few papers on on that refugee emergency I thought it was so wonderful and I was making arguments about for instance how the flows between spring and summer 2015 and early 2016 had changed in terms of their nationality composition or their gender composition or you know children versus adults etc but then I started asking my myself even the mystifying the very UNHCR and the kind of data they they they collect and what for what reasons but but particularly the also the IOM the problem of being in a transnational actually jurisdiction having no clear oversight and of course the idea that we needed to you know to make sure that we're not even unwillingly and unwittingly take advantage of the fact that people are in a vulnerable situation and okay there I mean again I'm not one of these people that say oh we're benefiting it disproportionately because we're making our careers on this data I feel that I'm trying to use it's a very political service in the wider sense by making my my you know doing my analysis and trying to highlight some some issues you know and and criticize policy and hopefully both shape public discourse and policy but but there are some very important issues because data unlike in the past unlike in the very recent past can travel so quickly and and so widely and they can even I mean there is such a big emphasis on how we collect and use this data and if we can go to the next slide so I want to I want to talk about then the no harm principle that certainly is very important and that's why I want to say this is something that we usually very much pay attention when we go out to qualitative interviews or you know ethnographic work fieldwork and so on and there is in most universities by now you know actually very oftentimes very bureaucratic ethics procedure on how you ensure that you're you're you're respecting the vulnerability of people giving informed consent what doesn't perform consent mean when people are not familiar you know with the written page that they have to sign they don't want to sign for many good reasons and actually I think it is it's a very important point to to to ensure ethics not as an administration thing but as a very important process that the that the research has to build particularly with populations that are in conditions of vulnerability so I think there are two two two things that I want to highlight so one is usually the police is you know knows much more than the researchers about issues that relate to criminal activity and even irregular movements nonetheless we need to be very very careful again it has to do with what kind of data and where do we store them I remember a few a few years back I had a great you know a collaborator who who was working in on the Central American migrant caravans that was I think in 2017 and we were discussing you know how he would make sure that you know he wouldn't put into danger any of the people that he was traveling with because he was traveling doing ethnographic anthropological work as part of the caravan and how to ensure that he would transmit the the data you know to me and I would store it and he would delete from his phone and all these issues so that was a very important thing and and and I think it remains a process rather than you know a procedure that you take boxes the other question that was pertaining to that research but also many times my own for instance research on irregular migration in Europe was how do when you publish your your results how do you make sure that the media coverage is fair that your your findings are not taken out of context and abused and they don't contribute to the you know further securitization of migration and asylum seeking and again I have in mind one particular occasion that was again back at the time of the Libya crisis and intensive border crossings through the Mediterranean and I had given them that interview to a journalist I trusted and I knew from a big French newspaper and everyone was asking at the time because I think it was the time when the Gaddafi had said oh there's millions here waiting to cross and the Italian government was trying to you know to how can I say refute this argument so he was asking what do you think and I was trying to explain that this given numbers is a bad idea because I said you know not not everyone wants to cross you cannot know and I was trying to explain also some things that I think Elizabeth was talking about how people make their decisions and I said you know this what they say about three four million is totally out of proportion even if we tried to make an estimate and I was explaining past data and I said something about 300,000 I think it was and I said but I was trying to to make sure that still this was not a number we could trust and the next day the title that he said it or as he explained to me gave to the the article was 300,000 I expected in Libya to cross Italy so this is something very important and I think with real-time data that we have it is even more an acute danger and here I'm not just talking about quantitative data like I was illustrating a minute ago but even qualitative data because those of these also travel in the last two years we've all been interviewed many more times than in the past to resume interviews because even the nature of journalists is changing these days going to the next slide and towards the you know the conclusion of my talk so I want to emphasize also in terms of the real-time data and the way they travel there is for the researchers obviously I think in the last few years we've had a few how can I say lessons learned about authoritarian regimes and police violence I think after the Arab Spring in particular I don't know how many people in this workshop are aware of the case of Giulio Reggiani a PhD student, an Italian citizen studying at Cambridge who was actually killed by Egyptian police he was doing research on trade unions in Egypt and I think this has sparkled the whole you know conversation a scrutiny that was already there but a much at a much higher level of awareness about when when researchers go in the field that there are important risks that need to be accounted and particularly that is true also for irregular migration for instance or as I'm seeking and again thinking of my colleague that was in in Central America with the caravan we were really concerned of course he was an experienced researcher but like unfortunately Mexican police are known not to treat very nicely people who come from Central America and you know are crossing with the caravan. On the other hand for instance another issue that is that merits attention is what do you do if you find out information that involves life-threatening risks and again for instance what do you do if you are doing research on migrant smuggling again I felt where I had done a lot of research in in relation to the Mediterranean and what if they tell you you know we're having this boat tomorrow night but it's not a sea-worthy boat but we get the money and you know we don't really care so this is also an important conversation that we need to have and again particularly given the technological you know advance that we have and how quickly we can transfer information and what are I mean obviously the basic rule of thumb is you try you tell people they shouldn't be telling you anything about criminal activities or illegal activities strictly speaking but of course if you if you find out some information that involves a risk of death you have to report their relevant authorities and going to the to the next slide so I'm trying to wrap up my my discussion I look also forward to to hear from from the other panelists so I'm trying to kind of come with some you know future directions I think it's not that we're going to go back to not having data not having them readily available or not transferring them so certainly we have to you know make the most of the the capacity that we have today but also think of our ethical obligations so I think big data and I think what again we have seen for instance in relation to environmental displacement is we can try and model and discover new relations by bringing together different types of data that are available and they can be available from different sources or not just population registers or you know data on you know temperatures or you know level of sea rise or or other environmental factors but also data that would not have been thought as immediately relevant like we heard just before from Elizabeth Fussell like you know social media data that are readily available and and can be processed or for instance you know cell phone data so I think we need more robust self-regulation at that point both in terms of the researchers but also in terms of the organizations and of the journalists and I think while regulation has to play a part it's also important to develop our own ethical codes I think self-regulation in this case is oftentimes a very important aspect because as we have seen with social media you can regulate to a certain extent but what is more more important we have seen in particular with the controversial cases of Facebook that having your own ethical code is is is perhaps more far-reaching now with regard to to quantitative data and how do we bring them together with big data how do we explore the dynamics of mess and micro factors and I think these are very important data because you know they help us make sense of connecting the dots or if we find some relations through a quantitative model how do we then further elaborate and understand the process of decision making but it is important to always keep in mind that our research can make vulnerable people relive very traumatic situations and it can also create false expectations and I'm sure all the people who have that policy work in this room have been with vulnerable populations have been in this situation where even despite what you say that you know you cannot you're not part of the decision makers you cannot influence decisions there will be hopes that you can speak about this particular person's case and perhaps help them say get forward with their asylum application or solve any other issue so I think there are some standard then challenges that remain that were there before the technology was more advanced but there are also some new challenges that really pertain to the great technological tools that we have today and I'd like to conclude with that and look forward to the other presentations in the discussion thank you very much and thank you so much that was terrific and I and myself have a lot of questions but I'm sure we'll be getting to that in the Q&A session and you really set us up for what we're going to be looking at in our panels this afternoon where we have a panel looking at the new technologies and the ethics and also mixed methods for analysis so thank you so much for that talk our next speaker into this first panel is Lydia Andrews from the Soldier Collective so Lydia I'm gonna have you take it away thank you thank you so much I am delighted to join you all from Sydney Australia where it's currently just after 3 a.m. the next day I'm calling you from the future and today I will be talking to you about co-designing alternative humanitarian futures so most of us who work in this space know that yeah protracted conflicts and unprecedented climate disasters are pushing the humanitarian system I'm talking about the international humanitarian architecture beyond its capacity and I'd like you to take a few seconds to think about your answer to this question that you see here just a moment of pause and dreaming so how would you reimagine the future of humanitarian action I'll just pause and give you a few seconds all right I'm sure you could use with a bit more time but other than daydreaming what else we'll cover today in the next 15 minutes I'll share a little bit about what it means to take a design approach I I'll also share a bit about a project with the humanitarian policy group which is part of ODI to reimagine humanitarian action taking a design approach and lastly I'll share with you where we're at with implementing one of those ideas that were generated from this project and it really focused on changing the humanitarian system to be more accountable and adaptable at still early stages but we can share where where things are at so first let's take a look at design and where it fits into the picture I think if we were in a room I'd ask for a show of hands as to what you know who's worked with designers or worked on design projects but let's um I'll just imagine um so you've got here scientists here on the left side of the spectrum also said to be more left brain and tend to observe the facts of the material world with an emphasis on quantities humanities professionals are said to be more right brain and tend to observe the complexities of human experience with an emphasis on qualities and so design as a third culture sits between the two poles of science and the humanities and design as I said to engage in more integrative thinking we are generally trained to synthesize both human experience and the facts of the material world in order to imagine and create objects interactions and value models that do not yet exist in the world so given this um design's emphasis is on appropriateness and it gets to the heart of questions related to feasibility viability and desirability when trying to create something new so now that we've established where a design approach fits in let's talk a little bit about what kinds of challenges a design approach can be applied to so you can see here um on the vertical axis a sketch of yeah a graph and um it's the degree of complexity and ambiguity of a challenge and on the horizontal axis is the degree of participation required for um I guess an optimal outcome um in a change process and um you've got yeah this amazing um professor Richard Buchanan who established the four orders of design to answer this question and the first order of design is dealing with 2d visual communication challenges the um the second order is dealing with 3d objects and materials um the third order is dealing with uh like four-dimensional interactions where time factors into the equation so think things like user experience of apps website and other services and the fourth order of design deals with systems so things like things like organizational strategies or social change and this is the challenge space that I mainly operate in and will be sort of covering today so when I talk about design that's where where we're headed um also just the last thing I'll sort of touch on regarding design um is how it's different to perhaps conventional ways of working um I guess designers we we deliberately design the extremes in mind this usually means that we also capture the needs of all users in a curve um to design sort of universally as we say um we also really place emphasis on understanding human experiences this is amazing quote like no one experiences the whole system we experience pathways through it um and so we dig into people's attitudes their expectations their behavior and this depth of understanding is what we use to then leverage um nudges and interventions that speak to people's innermost drivers um we also deliberately design in really highly collaborative ways um we um engage in extremely divergent thinking um before we sort of converge and and drive a bit of a testing mindset um and we have a bias towards action um failing forward and iterating and to learn by doing so that's sort of um a really quick summary for people around what you know what design what I mean by design when I'm going to talk talk talk about it um so in 2017 I worked with the humanitarian policy group and a hundred other people to reimagine what the humanitarian system could look like for the future and we I guess taking a co-design a collaborative design approach on this project started with building a very shared understanding of the problem space and what needs to change um you know I'm not going to go through that with you I'm sure you're all very familiar just to say we you know we identified sort of um these 10 key pathologies um and that was achieved through 75 interviews with aid recipients and practitioners across 23 locations and 73 organizations around the world um for each pathology we then outlined you know key insights and opportunities so for example that first line you know forgetting the human in humanitarian um you know the insight behind that was that international actors and either trained nor incentivized to be humble to really listen and emphasize and empathize I should say um and that has led to poor relationships and distrust and that the opportunity here is investing more in human to human relationships and trust building for connection so that's sort of an example um we also um to situate our design challenge initially we mapped the broad actor groups and the functions and the relationships through this wheel visualization you can see the broad actor groups um in the slices you know people affected by crisis government multilaterals etc etc um I guess that didn't really do enough for us this kind of mapping um from a from a design perspective we needed to sort of profile and personify the actors a bit more to be able to really get in their in their um minds and hearts so you know using we profile the actors based on composites of of those interviews that we conducted here you see them situated across a two by two matrix um it reveals the actor's group the actor group's relative capacity to influence change in the humanitarian system at present as well as their relative degree of effectiveness effectiveness as it relates to crisis um and so each each sort of each sort of actor groups you know presented um with a bit more detail which I won't go through today um what we found though was personifying the actors helped the co-design teams um to align and um and sort of you know align on on what's important rather than um and on who's important and help to humanize the problems we're trying to solve for rather than speaking about them in terms of statistics so um that was sort of humanizing the actors we also well I guess in this photo you see on the wall there two out of 11 um experience maps and they were based on human stories that were mapped very much in the first person and very much in verbatim form um and these stories you know were then annotated with barriers and enablers um which were posted by the co-designers during the co-design workshops and these were used as sort of platforms for change um yeah I guess what was most interesting I think some of these stories you can probably tell from their headings um but yeah I guess you know where people had found workarounds improvisations and uh unobvious ways to achieve their goals despite the current system and its pathologies rather than because of it and so once again they just really served as platforms for design inspiration um I'm going to read this out this is the future vision for people affected by crisis in a crisis that creates significant humanitarian needs every person affected has access to basic services safety and opportunity with the capacity to absorb shocks and the agency to shape her or his future and then for the system it was about a system that adapts to address the self-determined needs of people affected by crisis is built upon recognizing the agency of people communities and states and which can be held accountable to people for its failings I know these it can fit neatly on a on a slide today but that that was many many hours of dialogue and debate and discussion um we also talked a lot about this sort of you know ideal or preferred future experience pathway and here you know you can see it lays out some of the basic sort of needs or desires of people affected by crisis and how they may be met through many various touch points by very many many various channels so here it's about you know we have agency we are resilient we have protection we have assistance with community we have future we are self-reliant we have accountability nothing sort of surprising there but it was sort of thinking about you know how would we then design those touch points that make these things possible sort of shifting a little bit how we approach the action we also looked at the different roles I liked things that we sort of redefined here talking about storytelling functions talking about a multiplying function linking traditional non-traditional humanitarians custodians to safeguard and quality assure what's happening so I think that was really powerful as well there were hundreds of ideas that were generated but we did land on these 27 concepts that we developed in a lot more detail as well and sort of map them out and sort of process map them and how they actually work in practice some of them yeah you've got you can have a read you know failure targets having like refugee charter cities having plaintiff attorneys without borders having you know intended obsolescence incentives things like that we had yeah a community led response fund where communities can manage and allocate their own funds relief watch which you've got the loop sort of stamp on there because I'm going to speak to you about loop in a moment was you know initially an idea about having an independent watchdog which provides ratings of performance based on user and expert reviews for organizations and united beyond nations was interesting it was really sort of very anti-un in the in the dialogue in the debate but it was sort of bypassing it to be honest people connecting directly with responders and service providers and identifying that what they need directly on a digital and networked humanitarian almost like a marketplace platform type things are really interesting ideas so this brings me to the third part of what I'm going to speak to you today and that's where are things at with implementing one of these ideas and I say one but it's really it's turned into something that's sort of borrowed bits and pieces so you yeah the humanitarian system does not listen to people in crisis I think that we know that and feedback mechanisms and complete lines are traditionally owned by the organizations that are providing the assistance so usually donors are hearing and learning about the efficacy of programs they fund by organizations who are delivering those programs and so loop now exists and it is a safe and accessible platform for anybody to feedback on anything in any language without having to be asked that's that's it's it's goal so of course there are some brilliant M&E mechanisms and reporting structures in place to gather feedback from recipients of aid but as we know they tend to go to the same old people who kind of give the kind you know the feedback that organizations already know they're going to hear and I guess we're less likely to hear from more marginalized communities because of language or geographical barriers so with loop the idea is that feedback from people yeah you can also get feedback from people who didn't receive the aid and should have or should know and giving them a chance to find out why they didn't receive the aid or if the aid that was given to their communities was in fact the aid that was most appropriate so it's based on some principles you know decentralization open data dialogue being very proactive in terms of its accountability and it launched just last last year and is already operating in 14 languages across six countries it's very much locally governed and adapted to the local context through multiple channels which you can see there like you know channels that people already use already have access to so it's quite embedded that way I'll show you what you would see if yeah if you were someone who wanted to post a story using the web platform for example and there's there's also you know remember some of the research also hearing that that anonymity like some of these UN call centers don't really allow for anonymity because before you can give feedback you've got to give all your details first so yeah I think there was some really interesting design decisions made based on based on the conversations and the research that was conducted so also you know I'll let you have a very quick scan of one of these entries from earlier this week once received on the system a story is tagged so it's searchable and replyable and if contact details are provided and a particular organization is named there's actually the opportunity for the story poster to receive a direct reply which really closes that feedback loop you can also see here that there's a story type so story is a tag does either you know a thanks or a question or yeah an opinion a request to concern etc and there's also a whole different process for sensitive stories where the platform helps people safely manage and refer on issues such as sexual exploitation and abuse or protection or fraud and that kind of thing um actually a few months ago loop received a complaint about a human trafficking ring that was operating outside of a town in Zambia and actually helped get it yeah bust it basically so I thought yeah that was quite an interesting sort of use case that yeah there's some there's real sort of a line of questioning for us around you know how do we sort of scale its use for for those types of things as well um there are sort of gone through the stories side of the platform there's also the statistics side or more like analytics about those stories so maybe as a donor um they may not have the bandwidth necessarily to interact with each story or piece of feedback um but they could have the bandwidth to see okay well there's a lot of women in ex-province talking about shelter and sewerage after you know this storm or that event and um that's more of a high level sort of a way to interact with loop and allows you know donors and other decision makers to make macro decisions on how to allocate funds um and strategies while people who work in organizations um in country can respond individually to people's needs as they called on um so uh lastly I guess um um these are the countries that it's currently uh operating in it is a decentralized model um and it and it's very strong on having country-based governance and leadership and being hosted um by a national you know CSO and it operates a bit like a charitable franchise concept um and so yeah wherever it's hosted I guess is responsible to then you know train people and moderate the content in local languages and follow really strict moderate moderation protocols um we are still a little bit away from the end goal of being a platform for anybody anywhere to be able to feedback on anything in any language um but of what we you know what we do have is a platform where this is possible in Zambia the Philippines Indonesia and Somalia um and regardless of whether people have an internet connection or not people can feedback anything in any of the languages that they speak in those countries um some of these languages are also quite yeah underrepresented and marginalized in humanitarian space so I think that's yeah it's the first yeah first time for some of the languages to be digitized um there's a lot more I could say but I think we're probably um close to time so I will leave it there and I look forward to the discussions afterwards. Thank you very much Lydia and nice work being done at 3 a.m. in the morning uh really interesting um design aspects of all of this and I'm interested to hear some more in the conversation about what you're doing um but I'd like to move on to our final speaker in the first panel here who is three hours behind Eastern Standard Time uh Steve DeRoy who is the co-founder director and past president of the Firelight Group and founder of the annual Indigenous Mapping Workshop. Thank you Steve. Thank you very much and um can what are you seeing my note slide or we're seeing the slides it looks great it looks great okay I just wanted to confirm and uh thank you very much I'm really uh uh it's intimidating to hear uh such great speakers earlier on so hopefully I have something new to offer and so thank you very much for your time and uh and attention. My name is Steve DeRoy, I'm Anishinaabe from a place called Edmondflow Lake Manitoba First Nation but I uh call in from the Tlewatu Nation in North Vancouver British Columbia Canada. I'm going to be talking about Indigenous mapping today and how Indigenous mapping can be used to better understand issues around migration and uh just a brief introduction uh I own a company called the Firelight Group and we work for Indigenous groups across the country by providing community-based research and support and the way we do that is we equip staff with the necessary with the tools to be able to take on that work into into the future and and the Firelight provides a whole range of services to Indigenous communities in Canada um and so uh this just gives a quick overview of some of the work that we're involved in with Indigenous groups and uh and this is just kind of showing where our work has brought us to across the country so we're very fortunate to have built these relationships across the country um we're by because we're an Indigenous-owned company and the way we do our work is we really focus on the things that matter most to Indigenous communities and much of our work is designed to create and enhance community capacity and in a way that allow for Indigenous peoples to be in the driver's seat of most of the research that takes place and so the way we do that is that we train Indigenous peoples to actually carry out that research and be a part of the research team and so we and I'll talk a little bit later on about how we do that training and we have a pretty big workshop that we train people on how to do mapping but um uh this this kind of approach uh permeates across all of our business areas and all of our work that we do um so I love the idea of mapping I'm a cartographer by trade and so I love the idea of mapping and the power of maps and uh mapping has the ability to reinforce a relationship between space and place but that's rooted in symbolization generalization and um and in classification and and for Indigenous peoples maps have been used to assert power over territory uh and and the cartographer really holds a lot of power to decide what gets put onto the map and what gets consciously removed from the map and Indigenous peoples have constantly been moved um removed from the map and so although maps have played an important role for communication and for navigation the underlying notion of a map is that it's an exertion of power and knowledge and so those those cartographers holding the pen wield the power to define place and space and so and we've seen that with Indigenous peoples over time that have been constantly removed or displaced from the landscape to enable uh uh future settlement and so uh so a lot of the work that I've been involved in is is working with Indigenous peoples to use maps to decolonize and decolonize the map and uh and and move towards a new way of thinking which is actually indigenizing the map and so a lot of my work is uh working with communities that are in the driver's seat driving the process deciding what gets put onto the map or what not get not to map and maps are not necessarily new to Indigenous peoples we've been using them for generations to tell our stories and to assert our Indigenous rights and one of the earliest examples of maps that we've used are uh star maps and so uh by knowing the placement of the stars our ancestors could navigate to locations across the landscape and along the waterways and then some communities of star knowledge have been passed down through generations continue to be used today and so maps maps are just one tool in a toolbox to assert the rights and interests of Indigenous peoples I just want to make sure that uh we're not saying that this is the end of be all the mapping is the only thing but you have to think about um how else you might be able to tell that story and so uh for us it's about uh understanding what our rights are and understanding how we can apply them and for for Indigenous peoples the burden of proving our rights is on us really like uh we're not going to be relying on the governments or companies that are operating our backyards to ensure that our rights are acknowledged or enforced and it's important to understand the legal ramifications of where we're from and understanding what our rights are and understand how maps can be applied to ensure the rights of Indigenous peoples are upheld and so um as we've moved along in this storytelling and and data collection process we found that uh the technology has advanced so quickly and um and and the timelines for many of these decisions that are taking place on the landscape um are happening at such a rapid pace and so we've uh pioneered a direct to digital mapping method and where we project Google Earth up onto the wall and participants follow a uh a kind of a semi of um structured interview process we record that and we record the map data and points and I'm going to give you some examples of how that looks in a little bit but the idea that we could actually capture this information digitally and go through a major research exercise to then be able to use that information to be able to analyze and tell a story of landscape and migration and so um so we're in this place of of how do we uh re-story the map and so for these next few slides I'm going to provide you a variety ways in which Indigenous mapping takes place just to give you some context the Canadian government established the Indian Act in 1876 and that continues to this day and and there were uh the Indian Act was used to control most aspects of Indigenous peoples lives and it really focused on three main areas it focused on ban councils reserves and membership and its primary purpose was to control Indigenous peoples and assimilate them into Canada and so these points here on this slide really kind of talk about real elements of the Indian Act and so for example and we worked with some communities that have a lot of the Indigenous groups that we work with their lands have been expropriated for capital works projects such as agriculture roads railways and other public works um and I worked with a community that had been relocated not just for like a few uh blocks down the road but like from municipalities and regions like literally relocated from one municipality to another and they've been relocated about four times in their history and so this has major psychosocial impacts on Indigenous peoples and so before we carry out our research and when we before we even go into a community to do this work it's really important to understand this history so that we have a good solid understanding of of the of the impacts that colonization has and continues to have on Indigenous peoples to today um and and many uh parts of Canada Indigenous peoples still rely on boat navigation to access parts of their territory and boating is just one form of navigation that's used to be able to exercise uh Indigenous rights and so uh other other transportation modes include snowmobiles vehicles planes and helicopters dog teams walking through the bush um so but um in some of these examples that I want to share I think it's important to understand that these waterways are important for migration and understanding migration um so in 2011 and 12 I was doing a master's thesis where I was trying to understand the connectivity of traditional land use values of a First Nation in northern Canada and they had these this great information about where people hunt fish trap collect resources out on the land and and it was pretty much a map of just a bunch of dots all over the all over the landscape and what one of the one of the issues that I had with this map is that it didn't connect the dots it we didn't we couldn't see the connections of these places and so um I built a network analysis to be able to understand what those navigation routes could be it was a multimodal model that that looked at how to how these places connected to their to their community and then I went a further step to better understand well what happens if you have a certain industrial development that falls within the territory how much land could be affected and and and how much land could be alienated from a community for them to be able to exercise their rights in their homelands and so this really kind of set me up to better understand how we carry out the research and with indigenous groups and so um so we this is one of hundreds of studies that we've done but I just wanted to use this as an example because it is in the public domain and it's one that I can share and it was with the Tlicho nation up in the northwest territories they had a proposed mine called the fortune minerals it was a polymetallic mine called the nickel project and they wanted to build this mine in the territory near Tlicho and so we carried out 31 mapping interviews with community members from four of the Tlicho communities and we and we really tried to ensure that we followed rigorous research methods we obtained and implemented free prior and informed consent our data data we we were ensuring that that we were following OCAP principles for data meaning that the data I don't know if people understand what OCAP it is it stands for ownership control access and possession and that the information belongs to the indigenous group that we're working with and so um so we tried to establish a strong data governance model for this for this project in all of our projects where we use it to collect the information but at the end of the study there's a data repatriation process uh that where we then give all the information back to the community and firelight does not own this data um the result of the of the mapping interviews highlighted some really key information about about how this mine might have an impact on the on the nation um and and what we found was that the main water transportation corridor is known as the EDOT trail and that there were multiple accounts of critical travel routes and critical modes of accessing those surrounding lands in those areas um and and many many people were actually using those areas uh for trapping where they would you know follow these uh traplines on a regular basis and if this mine were to go through it would actually affect those rights of the of that nation so this is one way that you can go through a research exercise to be able to capture this information and really understand how that migration can be affected uh by uh by a proposed development um we also carried out a study called as long as the rivers flow and it was with the Athabasca Chippewan First Nation and the Mikosukri First Nation in northern Alberta uh and it was based on uh community uh knowledge of the river and um what the end result was is that um because the oil sands use a significant amount of water to basically separate the the oil from the sand they call it oil sands because it's like a sandy oily kind of sludge um they have to boil it using for every one barrel of oil they use about four barrels of water to be able to have that separation process and so um as a result a lot of these oil sands projects were along the Athabasca river just taking water out of the river uh to be able to do this uh uh separation and so what what happened was is that because of that the water levels were decreasing at such an extreme rate and and this research that we carried out uh we were able to identify two two elements one is the Aboriginal base flow um which identifies where treaty and Aboriginal rights with regard to navigation access and water level might be practiced fully and then the Aboriginal extreme flow which identifies where flow levels are likely to result in widespread and extreme adverse effects on access to territories relied on for the practice of Aboriginal treaty rights and so we did this study with elders and knowledge holders uh uh from the two communities and then in addition to that the nation did their own independent studies where they were monitoring water levels of the river and and looking at water quality and turbidity and a number of other issues and and basically after five years of collecting that scientific base monitoring uh basically the science confirmed what the elders had been saying five years earlier so this is where we can start to think about merging uh western science with indigenous knowledge um and so a lot of my work is is looking at how we might remap the territory using an indigenous worldview and understanding of place as well and so I carried out a place named mapping projects with an indigenous group up in northern albert or northern manitoba and and the whole idea was how do we reclaim sovereignty of the territory by remapping using indigenous worldviews and understandings of place and so many of these places uh included rivers navigational markers landscape features and we were able to get the Amish Nabe uh name the syllabics and then the english translations and this became a really important tool for this nation to uh to reassert their sovereignty to their territory and um one of the final examples I wanted to talk about uh working with a group in northern british columbia where uh they were uh there was a big proposed dam being uh built by mpc hydro um where they were going to flood the this entire piece river valley and it was the third dam along the peace river uh and and many of the first nations it was in their core territory where they hunt and fish and carry out their uh rights and uh so we went through a massive research exercise to better understand that but one of the elements of that that we that was an outcome of this is that we wanted to capture the natural beauty and essence of the peace river valley prior to it potentially becoming flooded uh and being underwater and so uh so we worked with google and the nunwuddy stewardship society to um basically uh mount a google street view camera onto a boat and navigate up and down the river to capture uh imagery of the peace river so this was a really important exercise for the nation to better understand on the ground what does it look like and how to how do the nations identify to that um one of the last examples I want to talk about is really about how do we uh um and I don't want to say blend I changed the word from blending to braiding braiding indigenous knowledge and science on maps and the idea is to create a multiple layer acts uh atlas that we can better understand those spatial relationships um and the outcome of this work can really design a preferred future for the nation uh that integrates both the indigenous knowledge as well as the scientific data and and one of the examples I've I've been fortunate to be involved in was with uh uh a a fellow named Doug Averly who was uh kind of he pioneered a method called bioregional atlas mapping and so um this this mapping exercise resulted in a co-management agreement with bc parks and the model is being applied in several other provincial park areas where indigenous groups have interests and so this is the idea where we're moving beyond just looking at specific information but how do we how do we understand the spatial relationships and how things are related to each other and so I think this is an important element for many nations as they move forward um but I'm noticing the time the pace of technological developments is staggering and daunting and for someone who's new to mapping where can they get started and we've really seen this transition from paper mapping to desktop community computing to mobile mapping to cloud-based gis and um many nations uh uh our a lot of nations are doing mapping but there's some that aren't and so how do they get started and so um so we've uh um started uh in 2014 the indigenous mapping workshop and for far too long indigenous peoples have been excluded from the map so we're trying to change that and and the way we're doing that is we're trying to build uh spatial literacy in the indigenous communities by providing access to geospatial tools and culturally relevant training and we've actually trained over 2,500 participants globally on how to use geospatial tools and we've had attendees from over 35 countries in all six continents and so we're very fortunate to have been able to build a global dialogue on indigenous mapping and and since 2014 the indigenous mapping workshop workshop has been an indigenous led event and it's focused on building a global community of indigenous mappers and so um so we're really uh these are some of the goals in which we're trying to achieve that that uh that collective and um and these are some of the places that we've done uh our workshops so since since we began IMW we've hosted a number of in-person events before the pandemic both here in Canada uh as well as in Aoteora and Australia and so we've supported our indigenous partners in those two places to launch those those indigenous mapping workshops and so um we're really excited uh for our future events we do have one that's going to be coming up and that will be um in Alberta in November so we're looking at hosting a workshop in Edmonton so uh look to indigenousmaps.com if you want to learn more about this um the other aspect of all this is that it's we've built this collective and and since the pandemic we had to move all of our training to an online space and so um so it was created in 2020 we've had we have over uh 1500 members uh on the on the collective and uh it's free for indigenous peoples uh if you're an academic or work for government or uh work for industry then um we ask that you pay a nominal fee to be able to access all of these training materials and this enables us to be able to ensure that we can continue to offer it for free for indigenous peoples. We have over 150 sessions that are available on demand we have over 100 hours of content that's been recorded and and put on on the collective and we have over 40 mapping trainers that are active and supporting this global dialogue so um the idea is is that we wanted to have a virtual platform that allowed indigenous members mappers to connect with each other and have premier access to mapping resources and for the first time mappers can revisit all the indigenous mapping workshop content at their own pace um I'm I'm out of time uh and I just wanted to say Chimikwetch thank you so much for your attention and for listening and uh and and I I think we'll switch to uh to uh thank you so much Steve that that's excellent um you're doing really great work and thank you so much for all of the the great information on how you're communicating and collecting this and mapping all of this information uh thank you to all three of our panelists uh Anna you know you did a really interesting uh talk about giving you know talking about real-time ethics how data travels so fast uh Lydia bringing in the design aspects of this uh looking at loop and ways to have safe and accessible information and then Steve also you know again just uh looking at integrating and braiding all of this I like that term braiding instead of blending this information with the historic um stories that we have with the indigenous population and then mapping that to make decisions um I'd like to open now for the committee members who might have questions for our panelists and for those of you in the uh public audience if you would put your questions into Slido we can try and uh integrate all of this we're going to extend just a little bit till about 205 so we do have some time for Q&A so I'd like to just pause for a minute and see if we do have questions from any of the um committee members I I have just a comment that I think of our our three presentations are very nicely complementary because I think I tried to signal some of the potentials but also the the pitfalls about the problems of technologies and I think what uh Lydia and Steve have done is to show also the other how do you use data what are again the what are the new potentials but also the limitations and I think it's uh yeah it's very exciting also the interface between you know the the subjects of the research the researcher uh you know the outreach and you know the or if you want the loop back to the communities it's fascinating and yeah lots of things to think about yeah and one of the things that you know I was thinking about when you were talking about you know this information in these different jurisdictions are collecting this data and in particular when you said you know some of the support that are given to some refugees are depending on the data that they're giving back and you know how do you trust that it's almost with you know Lydia and Steve you you already have some of that trust within those populations and so you know I thought that was quite interesting when you start to look at uh you know Lydia especially the loop I think that was very fascinating to say okay we have this in your your language we know it's accessible we know it's safe whereas when you're in a condition that you described Anna where you're you know in a flux and you know how do we know who we can trust I think that's you know quite interesting how you can set up those um I guess they're design ways or design communication ways to start to have trust into all of this um I do want to take a pause I'm not seeing questions but we do have questions from the public if you guys don't mind I might jump into those and then committee members we can jump back into your questions as well because I don't see hands up right now um so one of the questions we have from the public is have there been any use of citizen science and community-based dyi efforts and in context and I guess I think a little bit into that uh talk about again you know how do we get the community um you know citizen science engagement in this does anybody have any thoughts on that and I'm sure well through the indigenous mapping collective the whole idea is is that we have indigenous peoples leading their own research and really designing their own mapping approaches and so it is relying on elements of citizen science um and and by bringing people together and showing where you can access data how you might be able to carry out your research with the methods uh for various approaches to um research uh and here are some tools that you might be able to go out and do this it's it's definitely rooted in citizen science in a way that allow for indigenous groups to start building their own maps and telling their own stories yeah I think it's very different if there is a community so so I think in yeah again I I'm not sure but for instance in the case of um in some um in some of the work that you illustrate Steve um so for instance if if I think of how indigenous data were were collected certainly the community existed was there but it wasn't it didn't have the power so it could not um at the time you know um react or or was very limited or suppressed oppressed now for instance if we're talking about the population so even thinking about environmental disaster people fleeing or the cases where I was speaking say Libya you know um you know being the well not dismantled but really being in a lawless situation in the early 2010s when people live in these people cannot be conceptualized as the community because they will come I mean they might or might not have been so I think this is one element in terms of the agency of of who is being studied and whose data are we using that is important the other thing I want to say I'm not saying and as I said in this particular case of the displacement matrix we're looking into it and I want to clarify I'm not um as I said we're looking into it it's it wasn't readily available it wasn't easy to find so so we're digging deeper but I want to say so the the question is first of all that if you're fleeing and and there is someone providing for some support you you might not even think should I give my data should I not give my data you don't even have the luxury of thinking of that um the next question that I think is well maybe that that was done before but it was done by paper and pencil and it wasn't perfect as data the problem now is the data travels so fast I think this is both the big power and the big liability in there yeah Anna you know I couldn't help but think that in your cases you don't necessarily have the luxury to set up a design you know process for all of this because you don't know who those populations might be and um so I think that that's you know that's it's really interesting and you're right that data is is collected and it does travel fast um I see Pat has her hand up Pat yeah thank thank all three of you for really interesting talks I was really excited to hear about all of this ground up um impacted people generated data becoming available and the the clever and and exciting ways that you found to to get it and use it and evaluate it and it made me think about the contrast between these ground up direct but more informal data sources versus what we thought of as our formal data you know data generated by the government or the colonizers and traditionally we thought of those as two different data sources and I wonder how much the ground up people originated kind of data that you all are talking about is being accepted by the official data sources is it working do we really have communication between the official and quote less official but more original and grounded data sources or are the official data sources still resistant to to looking at these other kinds of data have we gotten past that barrier I can jump in here on on the loop front um we haven't passed that barrier and um when I spoke there's a there's another designer who leads loop I sort of led the sort of strategic work and he's more the sort of interaction design um the platform design and um I spoke to him two days ago and he said it's still really cold out there um and um I think they were they had a meeting with some donors who are funding the platform I won't name who but um and also some sort of legacy multilaterals um and the donors had convened this meeting because the multilaterals um were really against uh loop sort of being developed and and sort of you know increasing its um its geographic scope and its language scope and they the the resistance the claimed resistance was based on security concerns um and that's all been designed for to a tee like it's been it's it's been done really really well um and so that was you know explained um in detail um and there was still resistance um so I think the donors now have sort of come back to people you know looking after loop saying it's clearly political people are feeling threatened that their sort of business model um is under threat and their sort of sense of place in the world is under threat um and so yeah I think I think it's a really important question I think when we were doing the research initially for loop like this you know the very very sort of early stage research um we had partnered with a um an organization I can't name either um that does a lot of survey work and uh midway through our one year of prototyping they decided they didn't want to have anything to do with the project anymore um because it was too qualitative and we'll you know we were putting too much faith into data that you know was kind of you know too it's yeah it didn't have integrity and um and I see that as an extremely colonial perspective that because you know you're white and you sit in the global north that the data that you make sense of um has value to you but because it's someone else's data that that's not um you know that doesn't you know fit that normal you know normal or you know traditional sort of um way of of you know the world views um that yeah I think you know what I'm trying to say it just doesn't fit therefore it's invalid or it's not doesn't have integrity um so and so it was quite confrontational and and challenging and so I don't think we're there yet there's still a lot of work to be done yeah work to be done yeah um Marguerite do you have a question uh thank you yes um my question is for um Steve DeRoy but it could be anybody I was interested in generational differences of the acceptance of the technology and also maybe the trust or distrust of who's going to use the data where is it going to do it is it ultimately going to harm us um Steve do you have different versions of your workshop for maybe younger users or older elder users that's a great question and and further to Pat's uh I was going to provide a somewhat of an answer so I'm glad you've asked this question Marguerite because it kind of answers both um the idea of integrating and accepting the data I had a few points one is that indigenous groups have been omitted from past data gathering efforts and it might have something to do with the simulation tactics of the government to indigenous data gathered is held closely to indigenous groups because there's been a history of misinterpreting and misusing that information and three the methods for data collection have been court tested and peer reviewed by uh lawyers and law firms across the country so how we collect the data has been uh has been tested but how that information gets shared is a different question and whether it gets integrated in national datasets that likely won't happen but to answer your question Marguerite about how the audience we we train everyone and so we're very keen on trying to build uh spatial literacy across the board and so we have people that are young people that are coming to the workshops we have elders that are coming to the workshops we have staff that are working for those nations we have decision makers and leaders that are a part of that workshop um and the ideas is that we we kind of design the workshop with um two tiers in mind we have people that are just coming into the industry that want to learn how they might be able to collect data analyze it and tell a story through a map and then we have an advanced streamed people that have been doing this for years and they're looking for advanced methods and advanced approaches to how they might do mapping and so we kind of think of those kind of we don't think of it more on a we don't want to be ages in that sense but we want to think about more of if you're a beginner start here if you're someone who's been doing mapping for some time you might want to start with these there okay thank you yeah thank you and you know Steve you know that's this whole thing with the community engagement and helping the populations to navigate and I think that's one of the things that you do so well is helping them navigate the government um you know collaboration and intervention um you know unfortunately we're out of time but I do want to say that there were two other questions for us to maybe think about for the panelists coming up and one is a question about the range of non-human data that can be used as proxies for the human movement um you know things like diapers or pet food and you know other other supplies that people need as they're moving and that are not tracking the people themselves but the things that that go with them and and I think that at least that talks a little bit about what you had been talking about earlier and then also what types of ethical oversight might make sense for a programmatic data collection and again this is you know human subjects research data collection and we're going to be talking about a lot of that in our next two panels coming up so with that I thank you again they were as you said Anna of you know three diverse very interesting talks Lydia you know design thinking I think it's so important and what you guys are doing and the Sonder collective um and engaging the citizens from you know many different countries and Steve your focus on the indigenous population is terrific so we're going to take a quick break you now have 12 minutes instead of 15 minutes to maybe answer one last email on the break and we'll be coming back at 2 15 when we have our second panel which is new technologies for tracking human movement and the ethics of using them so with that we'll take a quick break thank you everybody good afternoon and welcome back to our workshop um we are heading into panel two my name is budu Bhaduri and I am a member of the geographical sciences committee um on the panel two new technologies for tracking human movement and ethics for using them uh we will hear from three very exciting speakers of our new technologies on tracking human movement and the ethical concerns that we have associated with using those technologies each speaker is going to speak for about 15 minutes and we will hold questions from the committees and those are few listening in through the webcast until we have heard from all the panelists with that our first speaker is Dr Miguel Roman who is a senior director and chief scientist at Lido's so Miguel the stage is yours now thank you good afternoon everyone from the east coast I would like to start by talking about um and presenting a you know a couple of things a human ecosystem and satellite based perspective on migration and display displacements I am I enjoyed our pre-meeting with budu a few days ago which some of you may not know but we're the academy is very organized we have pre-meetings and I was going to do my classical NASA scientists spills about how wonderful satellites are for fixing everything you know because technology like satellites are a panacea they're the solution to everybody's problems but I think I owe it to budu and the team of the academies to actually think more critically about technology and so I'm going to try to play multiple hats here so let's go to the next slide not just us as an individual that grew in Puerto Rico who has dominic and blog in his brain in his in his veins someone that grew up dancing merengue and salsa and someone who used to sing to the tune of Juan Luis Guerra in 440 there's this beautiful song called visa para un sueño visa for a dream I'm trying to translate guerras you know music which hits into some of the notes of why our dominican brothers and sister sisters have gone through the struggle and you know often violent struggle to pass through the monas passage and become members of productive members of Puerto Rican society and so I also want to so I want to hopefully tell you that story talked about some frameworks talked about some data and then we'll end and hopefully go back to from where we started I really enjoyed today's keynote and I think hopefully this will build upon the existing frameworks and understanding of displacement particularly as we think about the use of technology and society and nature essential to our helping understand displacement and migration so let's begin with nature next a lot of people may have already seen this picture of the 2017 hurricane season and we all have lived experiences around particular events that change our lives and change our professional mindset around the issue of migration development you name it and my case is Hurricane Maria and hopefully our keynote speaker already discussed beautifully the implications for for today's discussion I want to add a little bit more color to my experience and I think in spite of it being a little bit repetitive it's important as I tell my students that we have a critical view and learn the important lessons of previous events no matter how traumatic or painful that may be and so let's start a little bit to talk about hurricane Maria in terms of the causes of displacement if you go next there are some numbers these numbers are very inconsistent from study to study some estimates talk about 100 000 persons displaced among others over 200 000 same thing with mortality rates excess mortality the numbers go from 30 official numbers to 4 000 as statistically driven numbers from the Institute of Statistics in Puerto Rico that was the kiosher paper but what is what I care about is what is the oh no which is the emotional impact that hurricane Maria had on the lives of those living in Puerto Rico who are not just only Puerto Ricans but also include many members of our Caribbean diaspora Dominicans not only Dominicans but also the refugees that we took in prior to hurricane Maria when hurricane Irma destroyed Barbuda and we had to evacuate 2000 people to receive emergency medical assistance in Puerto Rico and then Hurricane Maria impacted the island I'm going to go and present a framework that hopefully provides some clarity in the importance of integrating technology humanity and society and hopefully this framework can be used and operationalized in a data-driven perspective and I'll introduce only one dataset that can help us get further in that discussion all right so I'll go next we already talked about the issue of livelihoods in a matter of hours the entire agricultural community in Puerto Rico collapsed after years and years of developing cells of development of agriculture in Puerto Rico so that's one key sector and cost of these placements next the other one is mobility this is the the bridge from the town over about 800 families live on the other side of this bridge and had to wait six months for the federal government to provide an emergency bridge and assistance relating to a complete collapse and of you know that is directly between remoteness and the issue of displacement and finally I want to talk about legacy conditions so the next slide this is a picture taken two weeks ago there are over 3,500 occupied housing units in Puerto Rico that still have blue tires instead of roots to to protect shelters of families and they have been recent close calls with other severe rents like Dorian that have already exposed how ill-prepared the island is to withstand additional storms and so all this is allows us to then frame the importance of today's discussion around the quality of interventions and so you know I'm a member of the round table on sustainability in natural economy so sustainability is something that I care a lot about so the next slide and I'll recording in progress is the issue of the quality of interventions on the lab you see the emergency food supplies that were provided by the federal government across rural communities in Puerto Rico and I'm gonna just quickly read through this go every time I see this picture I went um since Hurricane Maria nearly half of all citizens of Puerto Rico are suffering from hunger um and partly because of the recurrent acute disaster of COVID which merits its own round table shipping delays have increased the amount of time food imports arrive on the island uh the shale stable means that you see on the right included baby roofs airheads and expired cheese at crackers not to be surprised from a country that classifies ketchup and french fries as fresh vegetables the one on the right was provided by local NGOs and communities that were displaced within the San Juan metro area many of them Dominican immigrant communities they included you know very comforting food and brands like adobo boyo arroz rico a dozen eggs pasta imported from the US mainland and local resource canned meats and shell stable milk cartons one of the unintended legacy positive legacy conditions of Hurricane Maria is that it is it is making people think how we feed our people that are displaced um the work center kitchen which is widely known for their intervention across major conflict areas started um its work in Puerto Rico after Maria we just had a discussion with them uh two weeks ago at the academies all right so all this is a big mess right we have an issue of displacement that is tied to key sectors and these sectors are tied to our larger human ecosystem model uh that we are now presenting um and recently published so if you go on the next slide uh with members of the academies and those who are look at this from an ecological systems perspective um and so this is the work that uh we introduce is called recurrent acute disasters and it uses the work of uh bill birch gary maclus and joyland force the human ecosystem model which has been tested um in crisis settings including the big water horizon old spill and hurricane sandy so if you go on next let's talk a little a little bit uh about the birch model a good example of what we're talking about today is the interactions of multiple systems we look we look about flows of individuals from one place to the next and I think we need to start thinking more in terms of what these individuals provide in that as they flow from one human ecosystem to the next for example in Puerto Rico we only the only there are only two remaining brain surgeons in the whole island and so and that is a service provided to the entire eastern Caribbean basin you need to come in and be dropped up by helicopter if you're suffering from a brain aneurysm same thing with particular diseases same thing with oncological diseases for pediatric cancer and so when we think about this placement we need to think about the brain drain that oftentimes uh afflict people in terms of duration because it's it's a generational brain there I am you know I am a member of the Puerto we can die after in the DC area so I can tell this uh out of personal experience but there's also as we talked about in the previous session a flow of information a flow of capital um that it also needs to be tied into the flow of people and all this needs to be rethought about in the context of a changing climate where we're seeing the increase in recurrence frequency failure this is the and the expansiveness of new disasters I I was thinking about you know I joined at 12 p.m. so I was listening and I thought wow when we talked about force or plan retreat can we really think about force and plan retreat under a changing climate because people are just gonna be jumping around from one place to the next Puerto Rican moving from you know Dominicans going to Puerto Rico and then moving because of Maria to Miami which is over coming through sea level rise so where they end up in Texas where they get you know um freeze freezing uh that led to a lot of people to be coming at risk as well there is no corner in this place that's not at risk in our planet and so so that that freaks me out and so I think about one particular measure that I think that is central to understanding this displacement and I want to only focus on one electricity and so if you go next electricity is so central to the life, livelihood and well-being of a society those that are displaced require electricity oftentimes if they are more vulnerable either because they have need access to refrigeration for their insulin or they need access to you know under a changing climate air conditioning equipment so that they're not exposed to heat waves it is the one aspect that I think we can treat it as a keystone species in the same way we do with human ecology if you remove this function this basic function then sustainability collapses and human suffering accelerates and so if you go next what we look into is okay solution so now now I do my pitch so satellites so we have earth at night technology that we've been developing for 12 years and it's not perfect but it gives us at least a proxy of human conditions from space systematically in near-world time at a sufficiently spatial resolution even finer than track level so that we can start piecing the human ecosystem together with other sources of data and so if you go next here's some examples of one two three four five six examples of how in near-world time satellite derived nighttime lives from forwarding systems can be used to look at the causes and risk of displacement we can start with Aleppo after the Battle of Aleppo the entire destruction of the city we can see the flow of refugees into the Elzatari camp in Jordan and so people leave one place and add to the other we're just looking at the lights at one three in the morning however so we're tying to tying an infrastructural service to human population and we need to be very careful on how we do that COVID-19 lockdowns large-scale migration that is being led by jobs that are going away in places like Dubai obviously Hurricane Maria in Puerto Rico San Juan is used as a standard for comparing the rest of the island so there's always a means for us to anchor nighttime lives against other places this one's interesting Venezuela and the collapse of the Bolívar when there was hyperinflation in 2015 you can see this slow onset migration across all of South America and in that scene in the light and then traditionally you get seasonal migrations too because people move because they're looking for work and their goal where there's a lot of work all right so we go next let's go back to Puerto Rico this is social media data I want to acknowledge the authors of this paper Acosta and Miri Sahi who are well known at looking at really fine resolution modernized measurements of you know where are people at an individual point in time and I want to put attention to the blue change their active change of 75 percent in the mountain coffee region of Puerto Rico in the bottom and then let's compare that this is the important thing about being data scientists harmonization and cross-validation of measurements so we compare this against the lights next we can start looking at fine spatial degradation at 30 meter resolution within three hours so this data we didn't stop measuring this after 30 days after Hurricane Maria we are still measuring this to this day and so how does that look if you go and look at the municipality level and you can cross compare the results next you can start looking at urban areas comparing to rural areas confirming the results from Facebook that tells that rural communities are disproportionately displaced and more affected because of these keys relationship to this keystone factor electricity I asked my mother who stayed with us here in Puerto Rico when she was displayed when do you want to come back and every one of those 100,000 or 200,000 displaced refugees at the same thing until the power is back and so we need to start looking more broadly at technology and the technological failures that oftentimes are the result of a centralized system of response and inefficiency when rural communities require more informal assessments of risk in a displacement complex I will end by introducing one dilemma to today's panel if you go next and it's the dilemma that we're seeing a density vulnerability trade-off when it comes to when we start counting about changing urban patterns that are going to be happening 10, 20, 50 years from now and the dilemma is that who lives where can in turn shape the inequality that we see in forced displacement and and so you know as we've seen that we see that in China we see that in many places where we really need to pay this really important attention at the issue of scale because ultimately where you live ultimately affects the outcome from a migration context all right so I'll end it with a few points just reading the fact next that you know we need to extend analysis not replace extend and bring up bring to further analysis of population dynamics to assess the sustainability of societies and the sustainability of our biosphere to be able to withstand to recurrent events so that we can live in this planet we need to look at keystone variables not just energy but also food and water access to water we need to start thinking about the issue of frequency and I think what we're proposing is the characterization of legacy conditions this is very important because what happened after Hurricane Katrina absolutely changed the entire outcome and behavior of future disasters and we need to get better at characterizing that we need to get better at developing data and tools that measure legacy conditions and finally scale is a big importance but but you know it has to be driven by important societal need and we need to include the forgotten to advance sustainability sustainability for the sake of it you know doesn't help us we need to be very targeted about that that's all I have thank you thank you Miguel that was a fascinating presentation I appreciate you taking the time and sharing your personal story as well our second speaker is cascade to hold ski from columbia university he is a postdoctoral research scientist at the center for international art science information network at columbia cascade the stage is yours thank you thank you so much and thank you for having me I'm just honored to be here to present today and you know Miguel's presentation got me thinking I'm I'm a montana and I was reading the climate impact assessment for the greater Yellowstone region over the weekend before the flash floods hit could we go to my first slide please my title slide thank you and before the flash floods hit Montana this week and they don't mention flash floods as a climate risk for Montana and so Miguel's presentation really made me start thinking about just unanticipated risks and how both on the human socioeconomic side but also the climate side how we incorporate that into our decision-making frameworks um and really to his point about legacy um of these disaster events especially the ones that we don't anticipate happening um in places like Montana with flash floods that wipe out whole communities so anyways um I will be speaking today on understanding greater population datasets to measure demographic processes and hazards exposures worldwide next slide please and my talk is really centered on some top-down kind of blunt instrument data sets that are increasingly being used and I'm going to go into them in detail and great population data sets I think are really really important and if you take one thing away from my talk is it's just important to understand which granted data population data set you use and in what context because you'll see that they can provide very different information and that information can have real world outcomes especially when we're looking at a disaster response and um trying to anticipate the movement of people over space and time so um to to bring some context into it um you know for much of the planet we lack really fine-grained historical longitudinal data on human population um so this is a great paper uh figure right here from water up at all 2018 that's from PNAS and they just systematically lay out you know which country has had a census and when and how publicly available is that uh data and you can see for many of the most populated places on the planet we really don't have publicly available high resolution census data on top of that um United Nation population data estimates and our rates can be inaccurate often because they lack good underlying census data furthermore refugee camps and population data often are not provided or included in censuses this is especially important for IDPs or internally to place persons and so using traditional data sets like censuses can be challenging if we want to understand the movement of people across the planet next slide please so um this is just a very brief overview of gridded population data sets various teams around the world uh have started to augment census data with remote sense information to create a uniform surface across the surface of the planet where each grid cell has people in it so there's generally two broad monitoring approaches there's a bottom up and a top down the bottom up approach is to take a microsensors or area where we have really good census data or survey data and then build some spatial weights with earth observation data and then use an algorithm to go out and test the population throughout areas where we lack good information the top down approaches the inverse of that so we take the courses gray or the finest resolution census boundaries we have and then use satellite and sometimes GIS data to say here's where we think people are let's allocate the census data into those grid cells and there's I think four or five or maybe up to six or seven different teams who have built different gridded population data sets and different modeling approaches next slide please the difficulty here is deciding which gridded population data set to use in what context so the figure here is a map of Nepal and here we brought in five one kilometer resolution gridded population data sets and we just said for a given pixel which gridded population data set says there is at least one person in them and so you'll see the dark blue areas or areas of agreement across the data sets whereas the the light green areas you know only one or two of the good population data sets really says hey there are people in these places and so again from a very decisionary making framework in a hazard context or displacement situation depending on the data set being fed into the decision-making framework you can have different results next slide please so I'm going to present a few case studies right here is the 2015 earthquake from Nepal the primary the epicenter of the quake was really just north and northwest of Kathmandu officially two million people were displaced next slide please but if you look at here we're not measuring actual displacement but just exposure using these gridded population data sets that using this extreme event as kind of a case study to understand how these products differ you'll see or difference of one million people estimated exposed to an intensity of this is an intensity scale greater than seven so really the other thing that stands out to me is whether in a population designed as a rural impact or a broadly broad brush urban impact and so for really data sparse regions that are exposed to both climate and national hazards or complex situation these again these data sets are really important but they will provide different information next slide please and so in this this paper we walk through a couple of different hazards and also into some anticipatory action frameworks as well and we show is a similar finding and the important thing I want everyone to take away from this is that there is not necessarily a right data set but rather especially those like me who are scientists need to be clear that it's hard to quantify the uncertainty behind these products and that there are some products you know they're depending on how these gridded population data sets are modeled it's there's some clarity as to which data set you should be using in which use case and I'll get to that at the end of my talk next slide please so the same issue exists when we look at urbanization rates so this was a project I led as a graduate student where we were trying to understand urbanization for very broad brush urbanization rates using gridded population data sets across the continent of Africa and the first thing we noticed is just even the urban boundaries if you're just using raw density thresholds vary across these products sometimes it has to do with the satellite data that is fed into them sometimes it has to do with different census units being used to create the data set or when the census was taken but this makes it really challenging to measure flows of people using gridded population data sets over time from rural areas into urban areas if you have this kind of discrepancy again these products are not necessarily do not necessarily provide uncertain estimates or confidence although since this paper has come out those groups are starting to produce that auxiliary data for decision makers next slide please and so this is some more recent work by my colleague Dana Thompson where she and her colleagues decided to do small area estimates of slum or deprive or deprivation areas or low income areas of several large African cities and they showed that almost uniformly all of these gridded population data sets really under count the number of people officially living into in low income areas I believe this is from the figure right here is from Lagos and you can see that as these higher resolution products come out they're usually satellite drive you know down to 10 meters and they're estimating how many people live in a 10 meter grid cell that even those products don't necessarily reflect the situation on the ground and I'll talk about this a little bit at the end but I'm really worried that we're kind of trading accuracy for precision as we try to zoom in deeper and deeper without any sort of bottom up information to validate how many people are in a given area much less where those people go under a forced migrations scenario and this is especially important when we're using building footprints as a proxy for people again at these really small fine grain processes within say an urban area or a refugee camp next slide please so this is some work my colleague Jamon Bannon Hook is leading at Oregon State University to assess again these gridded population data sets for refugee camps and the first big takeaway from this is with official refugee camp locations often are not within 10 kilometers of any identifiable like visually identifiable settlement from image from satellite images so if we're feeding that information into a complex you know satellite coupled survey system that underlying training data may be highly inaccurate and we might be dumping people into places where they're not or more importantly not identifying people who need to be counted in these scenarios next slide please so this is all two again I use gridded population data sets in my research I think they serve a really important function for measuring human movement across the planet over space and time and they're increasingly being used in climate projections so this is the World Bank's grounds well report the baseline data this report are calibrated with an initial condition in 2010 both from a population grid cell population site gridded population standpoint as well as things like crop information and baseline climatic conditions and then basically climate changes push through those grid cells and then a gravity model distributes rural populations towards urban areas next slide please and with this they come up with a likelihood of climate in migration across the different I believe this was based on the cement five models but it might be cement six and they estimate across six major world regions the flow of about two hundred million people displaced because of climate change into urban areas or the caveat is that you know the initial gridded population data set they use is just a single one that as I've said differs from some of the other ones and as Miguel kind of alluded to they're under the assumption of a rural to urban migration framework but that doesn't really account for people bouncing around or you know urban to rural migration urban to urban migration patterns that these really acute climatic events may lead to the greatest movement of people very rapidly and then those people move to an area that's climate vulnerable and then they get hit with another climate event so I think really distilling out you know slow onset versus acute and just recognizing that at the end of the day we may not have all the information that we would like to have and being really forthright with you know what can we actually glean in a chaotic and compoundingly complex world we're living in next slide please so just some quick key takeaways and I'll pass this off if you want to learn more I really suggest going to this website www.popgrid.org there's an interactive feature where you can look at different grid cells from different areas it also just provides updated information on the different gridded population products as they're produced again I think these products are extremely important but the caveats of the uncertainty needs to be conveyed the easiest way to do that is just to read up on the given product that you're going to use and again with these increasingly high resolution products we need to understand we may be trading off you know this better precision for lower accuracy information and whether that actually aids us in making better decision I think we'll have to figure that out and then the last thing in terms of human population movement from my own research is that clouds are always a conundrum when we're using earth observational data to measure human processes so especially in the tropics in terms of real-time information we might not always have that even as these modeling frameworks get better so thank you so much for your time really appreciate it thank you casket that was wonderful that was terrific I'm sure that a lot of people are you know eager to ask questions but before we get to the qs session let me introduce our third and last speaker of this panel Caleb Litaro representing the GDL project so Caleb it's the stage is yours thank you so much for having me it's great to be here can we get my slides up well while the slides get loaded so I'm codley taro so the the the vision really of the GDL project is essentially this you know this essentially this this dream of scooping up the world's information and there we go all right thank you perfect so the vision of the GDL project is really this idea of scooping up the world's information and trying to catalog global events in real time and GDL today actually powers a large fraction of of actually of actually a very very large fraction of all global risk work done globally either directly through GDL data sets itself or indirectly through all of the myriad risk data sets around the world that consume GDL today next slide so the idea of GDL is really about how do we take news and other open information and transform it into data that allows us to really understand the world around us especially essentially taking if you think about news and all this other news academic literature social media etc is designed for human consumption uh it's it's the spoken word it's imagery it's moving imagery it's text but at the end of the day it's not codified it was never intended for machine consumption so how do we use machines to process that information make it codified but then how do we develop these analytic workflows and allow us to really start understanding the world from it next slide um so it all starts with news and this is actually a map of a couple of months of all the locations that GDL has gathered news from or about and you can pretty much it matches global population that essentially where people live on earth we're gathering information from and this is very unique because GDL today we have an enormous emphasis on local media print broadcast web in local local sources so this differs you know if you think about political science and the social sciences you know for the longest time it was if it's not in the new york times it didn't happen and still today so so much of that of the work is done in english or a handful of european languages and western sources in fact one of the major counterterrorism data or one of the major academic terrorism data sets that has sort of become the standard and used heavily by government which also is oftentimes used to understand kind of migration flows around that is based almost exclusively on english language american news outlets um which you know again really kind of shapes your really blinds you essentially to what's happening around the world next slide uh so in the aftermath of the 2004 ebola sorry 2014 ebola outbreak we looked back and we actually discovered we had actually seen the earliest of glimmers of that through local radio broadcasts in guinea unfortunately it was in french the time period we were only translating a fraction of that so by the end of 2014 uh we launched this mass machine translation infrastructure so you can imagine GDL today we're scooping up print broadcast web and broadcast at both radio and television uh from around the world uh we're we're scooping all that in we monitor today over 150 languages we know we actually monitor more uh but language detection tools kind of break down once you pass 150 and we're actually about to launch a 400 language detector for that we live translate absolutely everything we monitor in 65 of those languages soon to be more than 100 which allows us to reach really really locally so in the aftermath of ebola outbreak we launched this translation infrastructure fast forward to 10 p.m eastern time december 30th 2019 uh we picked up uh the a surge of local chinese language coverage in Wuhan china about a SARS-like viral pneumonia of unknown origin the following morning a company in canada called blue dots global which uh the vile surveillance firm um that uses our data they have these machine learning models built in our data they sent out a worldwide alert one of the very first worldwide alerts saying based on this data and based on these other models we expect the global pandemic will be coming by next month and of course uh you know so it was it's a really powerful kind of statement about why it's so so important if we want to understand the world to monitor local sources and local languages uh next slide we also have academic literature as part of this so a collaboration with the u.s. army a few years ago we want to look we basically from j story looked at all the academic literature they've collected on african middle east we looked at so if you're familiar with dt ic uh so u.s. government unclassed and declassed output um and then a lot of other uh sources there to look at context and and through this we actually codify out of so you can imagine taking all this literature and and extracting out computationally topics geographies bibliographics all kinds of metadata so you can go in and you can say um what are the food insecure take the food insecurity issues between these two groups that tend to involve mass migration events and who are sort of the academic uh experts that tend to be cited the most in that specific area um and how is that changing over time so this is actually deployed as a production tool was it was was the development of this really so that idea of let's say that we're going and we're actually trying to intervene who are kind of those those find experts globally but more importantly you can start looking at comparisons of what is the government and academic institutions like the national academies where they focusing versus what's the reality on the ground and there's a huge piece that a huge paper came behind that we actually showed this immense immense golf between what academia and the scientific institutions like nas study which tends to be where sort of funding priorities it really has to do with western policy um and sort of the the funding uh sort of funding trends at the moment instead of reality of where where events actually occurring on earth and typically what we find is the academic environment um and the policy community is about 10 years behind reality so this is huge huge implications that we try to understand events at the ground next slide um so you can imagine you know we're scooping up all this information from from around the world um you know what can you do with that so um with respect to the area of my great now again g-dolls deployed in in any topic on earth there there's a there's g-dolls deployed in some way shape or form but looking specifically at migration um typically one major use case is now casting and forecasting they're trying to understand global risk they're trying to understand you know where might uh you know where might instability come out um how might that manifest to what degree might migration uh play a role or be an outcome of that and where might uh where might those patterns go what are the stressors that they involve that they might cause as they migrate and that might affect their own migration patterns in as most specifically how are societies likely to react to that so looking at sort of the the global risk environment uh with respect to this and that can be sudden exhaust and a shock so that can be things like even natural disasters um you know say an earthquake which might not be immediately predictable you can however develop these forecasting models for if that does occur what are the likely outcomes within the structures of this particular society next slide um and more specifically so this is actually a fascinating model that esa that esa easo has come up with which looks specifically at risk forecasting uh for actually forced migration itself uh so a lot of sort of specialty migration uh models and specific models around the specific risk factors whether that's food and water access and so on next slide um but of course once migration occurs like once an event begins occurring um ghettos oftentimes used to map that in real time so there's actually a fascinating map that bbva came up with uh during the 2015 uh uh exodus and what's fascinating about this is he actually drew and actually matched um it actually matched eventually government statistics sort of caught up because again government statistics are so so slow um and so what this bright was a real-time look at where flows were occurring sort of inflows and outflows local stories and specifically local reaction to that that for example a small town somewhere a group of refugees showed up how are they reacting to that and how is that reaction changing because again if we look at say the ukraine crisis right now um you know looking at say local polish newspapers gives you a lot of interesting insights to kind of those early tale warnings of sort of the the european reaction to that next slide and we can also look at the mapping of narratives so narratives again one of the most important things if we think about things like forced migration um it's not you know satellites will tell you where people are moving uh things like cdr records the cell phone data records will allow mass scale mobility um at an agris scale and those are of course deployed by almost all ngos and governments today but that gives you certain things and cdr records are very there's a lot of nuance when you look at refugee migration specifically around um certainly you know their their applications to different parts of the world different types of migration patterns but most importantly what they tell you is kind of where people are at this moment what they don't tell you is the most important thing which is what is the information flows that are reaching migrants which has an immense impact on dis and misinformation uh that can feed them so both real information but also again there are a lot of actors um both well-meaning but also malicious actors whether private sector or state uh or criminal enterprises as well that attempt to direct migrants that attempt to use them uh increasing as a tool of conflict um so feeding them force forcibly feeding them false knowingly false information um with this intent to create disruption they're sort of becoming a form of irregular warfare in many areas of the world and so how do we understand what are those information streams that are arriving to them but then most importantly the domestic reaction society and one of the things i'll show in a moment um is how those play and one thing you can see with this is mapping narratives there's ability to go through and this particular example is actually drawn from climate change um which of course has a lot of impact on migration um but that ability to so this particular example here was let's take a collection of of a few months of climate change coverage just take all the people that are mentioned in public news coverage and map the interconnections to those and what you see in all those communities around those are the narrative communities how climate change is specifically being internalized i mean around the periphery of those are the reporters the journalists the the experts in those areas so it becomes a very powerful way to take these massive complex and very fluid narrative environments we can visualize them in ways that we can that we can actually message into uh next slide but also understanding how the media is covering it so this is uh this is a kind of a global look at um coverage essentially of refugee attention to refugees since 2017 you can see the ukraine crisis caused this immense intense coverage of refugees of migration but we can see how quickly almost immediately within days of the war beginning it faded away and it's continued to decline um but also we can see afghanistan didn't really cause nearly that amount so we can see the world's attention and in return funding and policy and other support um tends to flow of course through the pattern of um you know what sort of the what societies um specifically the large donor societies kind of care about in the world which has huge implications there uh next slide we can also see that media attention to this national attention to refugees and migration tends to peak with their own impact so this was a fascinating uh chart that erwin did in 2015 and they looked at each country when did they begin their domestic press begin talking about migration and so you can see each country um as as migrations of migration flows spread across europe um countries didn't say oh look at what's happening we need to start talking about this it wasn't until it reached their doorstep was it well that's not our problem is someone else's problem it's only when they reached their own doorstep that the media started covering that and that means that that critical missing period we can start preparing a population discussing uh what's occurring here uh the needs of these populations all that messaging environment that's lacking so sort of the average the average citizen doesn't even realize that there's a crisis until all of a sudden it's on their doorstep and they don't really have a chance to kind of internalize or adapt and and that can oftentimes um see the sort of a negative reaction to refugee flows and so there's a lot of sort of missed opportunities there next slide um but then also understand how they're framed so what you're seeing in this map here is when you look at media coverage of refugees what are the countries that are mentioned together um and so again this is not the reality this is how the world's media kind of views refugee movements across the world there's a lot and again a lot of this makes sense there's a lot of telling insights there um in terms of how we kind of the lenses and specifically the western lens through which migration is viewed throughout the world next slide um and also now this is an interesting one so this is actually specifically of natural disasters um so this is looking at the media coverage of a natural disaster and specifically the essentially the exponential decay curve of coverage you can imagine um of a particular area there's very low coverage and natural disaster occurs there's a surge in coverage last 72 hours and the decay um it's almost a perfect uh decay up to 14 days now what's interesting about this graph is of course media scholars have always known this but what's fascinating about this is we can observe stories that deviate from this so we can take any emerging story overlay it onto this and deviation so it's inorganic that an external actor is inorganically shaping the narrative of that story whether it's a natural disaster whether it's refugees whether it's another story that has huge implications that tells us that this is no longer an actual story about refugees this is something else entirely um so it means that some other actors attempting to shape this whether that's as a tool of conflict to so discord um or so on and that once you see those kind of deviations that allows your analysts then to really drill and understand what is the messaging environment around this and what are what are the things that we need to do to intervene potentially next slide um now this map this is actually part of an animation which i i'm not showing here but this is actually fascinating so this is scooping up each day all take every mention of paris across the world's media that we monitor each day how positive and negative was all the coverage of paris yesterday if you do that across the world um what you get is of course not how happy or sad people are within an area but how the world's portraying that now what's fascinating about that is when we looked at europe and we zoomed in you can see the you can see positivity you can see these tendrils of intense negativity start making their way through europe that was the backlash to the 2015 refugee exodus and you can see that you could actually see in real time sort of this wave of positivity and then this turning against it as those refugees became sort of the the outlet essentially the scapegoat for anything that was that was happening in society inflation crime etc suddenly it's all their fault and so you can see that um remotely through the data next slide um and we can also look at things globally so this is actually a fascinating one of sort of global positivity and negativity of worldwide media uh since 1979 prim broadcast web over a hundred languages and what we can see is you look here you look at the air of the web you can see at the web that sort of tone kind of moved around but it was the web era as online news florist that we've plunged um ever more steadily towards negativity and that has a lot to do with media and how you know papers and outlets now are competing globally and social media and virility but what it means at the end of the day is as steven pinker actually used this this graph in his book i mean if you look if you look more into this you see that the world isn't necessarily becoming a more negative place but the media is portraying it that way as they you know due to commercial pressure and what that means is the lens through which we see like the the lens through which we see say migration is no longer going to be as it might have been 30 or 40 years ago here these are real life human beings they're migrating here's their challenges here's what they need instead it now becomes an opportunity for essentially negativity of here's another bad story um and and again that back that sort of that backdrop of negativity more broadly this has a again more implications than information environment next slide um we can look at simple things like pronouns within media coverage and we can start looking at are the our refugees portrayed as us you know as as human beings like we are or is it those and they and them you can see these fracture points in the media at key points next slide um we can look at morality call out so we can look at how are they being portrayed are they being portrayed as um you know are they um is this a moral imperative to address them is that patriotic duty is this selfish duty to assist them um what are their driving factors and again these are not the reality but this is how it's portrayed to societies which again has immense impact um to the sustainment of policy and the the facilitation of resources next slide um and we can see this very clearly in things like ebola when the ebola outbreak occurred it got very little coverage until the first americans got it and if you look at the bottom graph there you can see that the tone of media coverage of covid is very negative to the first americans got and it became more and more positive um and that again that was the the sort of the westernest savior that oh don't worry now that westerners americans have it we're going to go and save the world now and that again has a huge implication to the lenses to which societies and in term policy decisions of funding resources um you know all these different factors play a huge role through that lens next slide we can also see the impact to which major events um so for example after after kurti's death um you can see the tone of media coverage towards refugees is very negative after his death we saw this wave of sympathy there um then that of course faded away so the ability to kind of measure that in real time of of of these events next slide um and this is a fascinating visualization uh that kathiebus did was this kind of looking at media coverage of refugees and putting a dot on the map when there was a positive mention there um so how are refugees being portrayed next slide um this is another fascinating researcher um and what she did was a really really fascinating map is she took this media because she basically took our live feed and looked for any mention of refugees are and then looked at what came after that so what is a refugee is it uh is a refugee is a bad person is it a stateless person is it someone that's that suffering is it someone that's created that's a criminal how is it being described and how does that vary across the world next slide um so she created these these fascinating maps to kind of show the main you know are they communities you know are they claimants are they basically a burden to society or are they part of our community there next slide and she did these these really fascinating maps we can kind of see word by word by word how are they described so understanding those narratives in real time next slide and then finally looking at um so this is actually this is specifically about covid but this ability to create these large narrative maps so this was actually an example done with covid where the question was um it was essentially a group of of allied governors where we know the major narratives around covid what are all the narratives that are coming out that we don't know right now so the ability to sort of map those out in real time next slide um and then finally looking at so this was actually the the map of covid actually as it first occurred as the earliest glimmers um restart data sets um this was actually those early early glimmers you can see that first initial pickup um that's when we saw that final wave there was two weeks later when the west of the world finally started coming to terms that maybe there's a problem here so again that importance of looking locally next slide um and so finally if you think about the gdo project it's really this idea of scooping up the world's information trying to process what's happening there is specifically with respect to um to migration it's really about this idea of understanding the world's risk factors of seeing both what might be happening here but then eventually you know sort of looking at both what are the immediate risk factors but then other things like exhaustion is shocks what are potential scenarios there as things come to a head measuring that in real time um and giving sort of those those prediction those those essentially those forecasts once things actually turn to action and and events begin to occur mapping those in real times as those turn eventually into migration mapping those out and then as those as those exits occur how what is the information uh environment that's reaching refugees what is the information environment that's reaching societies and how are they internalizing that so thank you very much i know i've covered a huge amount of ground here but hopefully it gives you a little bit of glimpse of again this this broader idea of how we can use these non-traditional data sets that will understand global events both long term um but then also again that narrative environment which at the end of the day you can use a satellite to watch to watch migration flows um you can use cdr's observational data all kinds of different um data to kind of say well look here where people are but without understanding the narrative environment around that you can't really understand how societies are going to react and most importantly what is the sustainment and the range of policy options that are available obviously in ukraine governments around the world are stepping up in a way that afghanistan they did not um and that has a lot to do with again how it's being framed how societies are being told these stories um and so again that the huge amounts of insights i think that can be gained from that thank you so much thank you kelev that was fascinating um i want to thank our three speakers miguel casket and kelev uh for your fascinating um presentations and insights um you know uh miguel did a fantastic job of laying out the background and some of the things that are associated um some of the things that stuck with me are these uh issues around food and nutrition for idps um and then the connection the whole connection to sustainability of nations and states um casket's presentation on you know measuring population especially you know this impacted and potentially displaced populations from disasters using different kinds of data you know gridded population data um that was really um very informative um i particularly liked the issue about um the community's trend of trying to be quantitative with building footprints and and he very rightfully pointed out that there is an issue about building volume which we still do not understand very well and kelev's um real um insights from how do you turn news into data and this qualitative assessments of the narratives that not only allows us to understand broad general patterns of you know idp movements but also the sentiments around those populations and and the populations surrounding them so with that i'm going to open it up for questions from the committee members um and if you have one please um raise your hand in zoom um if you would like to ask a question uh for those who are our public attendees watching the webcast please contribute your ideas into slido and we will work to bring those into our conversations um so let me go and see who would like to ask a question harvey okay um thank you very much for very interesting presentations i mean just really i have i have lots of questions but i guess i have one underlying and i think very important question so you talked to the panelists talked a lot about um nontritional data sources that they get at things that are difficult to measure um but when we make legal and policy decisions in government we have really strict standards for data authoritativeness you know for validity reliability how representative the data are how transparent the data generation processes how accountable it is i wonder if the panelists could comment on that how do we take these nontraditional data sources and make them good enough for you know authoritative enough that we can make um you know policy you know policy legal and and infrastructure investment decisions who wants to take that money i mean i can start i mean you know in the open source within say the ally government's open source intelligence has an 80 year history of driving policy decisions um you know the the original warning of pro harbor came from open sources um it was actually fibbuses or fbms is very first report um so there's a long there's a long heritage and a long history there of trying to understand media narratives trying to understand these types of open sources and understand that exact question how do we translate that to policy how do we understand the biases the the various pieces that play into that so the short answer is there's a long long history what we've lacked is the tools and technology to focus globally um strategically rather than merely tactical of kind of the the the issue of the moment so harry i'll i'll go next and i agree there is a policy resistance towards um sources um but i also think broadly if you think about policy i think about like i i put i will put a marriage sheen risk management in that in that box and let me tell you if if you give fema 60 percent accurate thing they'll take it because they don't have anything else and so there's there's some land going on in the pipeline we as humans hate uncertainty and don't like to make decisions about uncertainty but guess what that's the future the future is full of uncertainty and so we need to be more a scientist and and more open to talk about error bounds and data quality and traceability and all these other principles that would allow policy makers to understand that we've done our work to give you the best information that we can but forecasting is is as good as the data that we have and so here's some investments that you can make if you want to improve the confidence of these predictions for example night time night everybody's asleep at 1 30 a.m in the morning so guess what half of africa all these maps that you see on night time nights they're not that really useful because electricity decisions are being made such that you know turning off the lights to you know maintain energy storage and so we need to think more broadly as a science community of what investments we can make to have to have more optimal global remote sensing portfolios that are in service to the humanitarian community i'm hopeful that you know this group would help advance those causes so that we can tackle the issue of uncertainty but you know trashing trash out if you're not going to you need to spend in make investments and fund research on frameworks that allowed you to reduce those uncertainties if there if you don't do that you're not going to get good policy outcomes so it's up to the us to communicate that to the policy makers yeah i'll just add i mean uh miguel you you said very closely to what i was going to say and i i can't recall if i said this during my talk it's my brain's little fuzzy but zoom but invest in traditional i mean on the demography side of thing invest in the traditional demography methods we have to invest in our censuses we have to invest in having people on the ground and understanding um individual communities on a very human level and i i love remote sensing and i use it all the time in my research but as miguel said very well garbage in garbage out and and i really worry with machine learning uh and artificial intelligence becoming where i see funding going everywhere that those models are only going to be as good as your training data and if your training data on the human side of things isn't valid then your model's not going to give you the information you want so traditional investment in real social sciences i think is really important as we build out these technologies all right let's see if anybody has another question while people are making up their mind i wanted to ask one question which is you know in our title of our panel we use the word ethics and as we as you all discussed the different kinds of measurements and analysis techniques for data i wanted to ask all of you your perspectives on you know how ethics should be playing a role and and i can see this with with one observation which is the the work that i am involved with you know for many years is you know identifying human settlements on uh from satellite images um you can pick up these refugees and IDPs in very sensitive conflict zones and that data being if it becomes public becomes incredibly dangerous because you are putting human lives at stake of where people are taking refuge so it became an ethical debate about you know what is the right thing to do you know do you release those kinds of data sets for public good or is it something that we should protect so i'll open it up to the three of you to see you know from your perspective how would you characterize the the role of ethics i mean from a media perspective that's one of the nice things about looking at at news media is again that that openness there i will say that you know certainly with cdr's and social media data you know i've you know in our past life wrote extensively on the data ethics area um and you know it is one of those scary places where you know facebook for example has this disaster mapping program it makes it available um but there's a huge amount of concern there because a lot of ngos um you know even major ngos like the un um you know there's a lot of interaction with governments and repressive governments can get access to some of this information um you know and and there is a shift i mean you know whether it's pn you know pna s uh triple as the journals um you know the proceedings of national academy whether science magazine um now actually uh the apa as well as journals um all of you now have done shifts to say a lot of these traditional like the common rule doesn't apply to social media it doesn't apply to a lot of this sensitive data um and it's really fascinating uh when you you know kind of look at the ethical shift that's occurred and it used to be that a lot of these data sets um used to be things that were considered very sensitive um and you know now um you know it was it was kind of shocking when you know pna s said um you know the all those rules about manipulation of social media monitoring all those especially for understanding population scale we don't see those as as issues anymore so i think that landscape is changing dramatically and it's changing in a way that benefits academics looking for data sets for publications not with but ignoring the very real dangers to society and i think that is a market shift um that i've i've you know gotten statements from all the journals now of yes we've totally changed on that now we don't really see this at these ethical constraints the way we used to we have to prioritize publication we have to prioritize grant funding we have to minimize the ability of societies of citizens to protect themselves it's kind of a very interesting uh piece there that really is not getting a lot of discussion i think right big old casket so i like to frame the discussion of ethics and data in a humanitarian context in a more positive landscape because i see and i have witnessed how this data can also be used as a mechanism for transparency and accountability of those centralized systems that are supposed to be helping the forgotten when you show an image of the longest power outage in porto ripina and us history and you say where the people have to weigh the longest outages i think that's very powerful and it has ethical implications to put FEMA housing and urban development army core these agencies which us taxpayers pay to protect our communities to task the same can be said across the global south when we see all this helicopter humanitarian work going in you know feeding off of charities and then the work doesn't get done so creating mechanisms of accountability to track the effectiveness of electrification programs across africa is something is one example where there there are still some ethical dilemmas but you're you're trying to concentrate on on the so that the positive aspects absolutely true you know we have to be careful about issues of national security you know i as i join light as it's becoming a more i realize as i oh my god we've got a lot of stuff with this data but we also need to make sure that we offer solutions in terms of equity and justice great that's good uh booty i love what you said about our ability to detect populations who are vulnerable who with that information may lead to bad actions whether that's by government or other people and i it's something i think about all the time um when i contrast it with what mcgill said it from a ham humanitarian standpoint and i think i tend to be more glasses half full and a little naive as to how some of these data sets we produce are used um the second part of ethics that i've been thinking about is whether providing data human settlement data or hazard data or a climate forecast that we're not very certain about is better than providing no information at all in a humanitarian standpoint and i don't have a good answer for that but it is something i'm becoming more mindful of is as we as i produce more data sets and make more user friendly is commuting communicating that uncertainty but again whether putting in someone's hand will actually just lead to worse decisions or overconfidence in a decision-making framework and i don't have a good answer for that but it's definitely on my mind so i really appreciate this question bulu i'm sorry i would like to add the bad actors unfortunately there are more and newer bad actors who are figuring out ways to use this data in ways that impact uh the spacements and i'll say a lot because i don't care look at air bmb and the land grab that happens after disasters where they go in and say oh we're gonna get free rent to people yes but then your your pop overpopulating this air bmb in in places where land is extremely limited and now people cannot live there look what happened in barbuda where people are had to evacuate and when they came back their lands were taken away uh by commercial bad actors so i think we have to be you know if you're like i say very ethical about policing not just the traditional conflict driven actors but the the ones that are actually intensifying the inequalities in a displacement context you know and i'll add one piece of that you know our our approach has always been to work with local communities you know so much of this sort of data analysis is you know folks here in the u.s or europe using mass amounts of data and describing you know like taking side imagery and analyzing it and saying here are the refugee settlements that we're observing in this country here's what we're seeing um you know our philosophy has always been the opposite of working with communities and saying here's what we can observe about your community um what of that is useful to you what of that is sensitive to the general public and allowing them to make those informed decisions and i think that's an inversion of the academic world because again academia that is an academia but i think that's something that really has to maybe become sort of the future of when we think about this because at the end of the day these are real people's lives and if you know professor in america publishes this big paper saying hey look we found all these all these hidden settlements here and we've mapped out where all these people are going and that's an area where you know they can run right into to death i mean that that's a huge thing uh great um i don't see any other question so and we have about uh four minutes remaining so um i would like to solicit your last thoughts um on you know do you perceive any technologies that might be emerging um in in in the next you know five to ten years that would significantly you know impact or influence this particular area that's a good one um you know i honestly think that we have a technology backlog we are swimming in data and we're still like look at our nighttime nights we're still trying to make sense of new technologies and we're living you know at least on our sector in an embarrassment of riches when it comes to data the problem as you have stated and others have stated is that policy making isn't catching up to to these advances and you know like i couldn't understand i mean kyle kaleb put it perfectly you have you know the government way of doing it you know that other academic way of doing it we're so ahead of the curve in in being able to provide informed decision making and yet it hasn't scaled and i'll be very frank it hasn't even scaled in our decisions around the future of investing in science programs you look at the current science decadal survey you look at reports that i talked about i think um the cascade was talking about on the climate assessment they're they've already expired and so i think part of it is you know what do we do with what we have i i'm looking i i know dr karen sanger may say sometimes at nasa our our um our headlights are looking down they're not looking out because we have so many things happening great thank you miguel any last quick thoughts yeah very very quickly out i will say i'm most excited to use the technologies we already have them and and work with communities on the ground who may not benefit from them i mean one area that i working on more recently is just extreme heat and and population exposure and just working with local communities on proving weather forecasts because we already have those technologies and they're really good in some regions world and they're not so good in other regions world but we don't need to reinvent the wheel for a lot of these regions that don't have good weather forecasts so it's i know that's tangential from this conversation but and i'd say from from my perspective you know one of the challenges we have so much data today um the trick is what we lack is analysis um you know our government the intelligence agencies for example um you know if you look at say the cia museum um is filled with things to collect information there's not a lot of emphasis on what do we actually do with it when we have it we you know we kind of as governments we and academics we specialize in in acquiring these vast rows of data um but the end result is how do we actually what we lack is those those tools and we have actually a lot of the technologies it's the the methodological workflows to to leverage the tools and data that we have to make new insights from what we already have great so with that i wanted to thank Miguel casquette and Caleb all of you uh and pattern harvey for planning and designing this a wonderful panel this was fascinating one of the best i have um seen in recent times in terms of the depth and breadth and and the insights that you have shared so i'm going to turn it over i guess we are going into a break right now is that correct yeah so we we um we are going to go into a break um and then we reconvene in about 15 minutes at 345 for our last and not pleased and the most exciting panel on mixed methods analysis of migration displacements and human dynamics that and the panel will be connected by no other than our um wonderful colleague Elizabeth root so um see you everybody at 345 hi welcome back folks um we are about to start panel three um which is on mixed methods analysis of migration displacement and human dynamics in this next session we have three separate talks on using mixed methods to analyze migration and displacement and human dynamics and again as with previous panels each talk will be about 15 minutes and we'll hold questions from the committee and all those of you who are listening through the webcast um until we've heard from all three of our panelists today i'm fairly excited about this uh session we had a great meeting last week to sort of touch base uh with the panelists and it's clear that we have a really interesting mix of different ways in which people are measuring and visualizing and understanding human mobility um and it offers a great diversity um across a different space and time scales and whatnot so with that our first uh speaker is uh samaya dodge from the university of california santa barbara so samaya i will hand it over to you thanks elizabeth uh let me share my screen do you see my presentation next slide yes it looks great great well thank you so much for having me i'm very excited to present a portion of my work regarding to mobility and responses to cove specifically as a environmental disruption to our daily life um um it was a fantastic first um a few panels and i really enjoyed the conversations and the discussion and i think what i'm presenting today really relates to the discussion we had right after the last panel about data quality and representative mess um so starting with the pandemic um a lot of companies and organizations started to share um mobility data at different scales especially and temporally um these data often we are calculated based on cell phone traces like if you are using cell phone and we have location services on uh we've been counted so on the left lower part animation you see an animation that was created by new york times about raw cell phone traces obtained uh by cubic company and they showed that uh you can learn a lot about these people although they are anonymous traces although not always we have access to gps traces for good reasons because of privacy issues but these traces are used to calculate aggregate mobility indicators that indicate how much for example people move in general and how much or how many people are moving so a lot of companies uh like a cubic the scarlet slab safe graph facebook apple map box started sharing this data with academic community to basically explore the patterns that we see in the data and this data has been used a lot uh with uh during covid for different studies like looking at the impact of non-pharmaceutical uh interventions such as stay at home orders on the spread of covid or modeling covid itself and also it has been used to inform policy uh in terms of making um uh non-pharmaceutical intervention policies so there has been a lot of studies and publications in the area but oftentimes these studies focus on one um particular data sets or use a data set or two sources of data to make these analysis and modeling what we wanted to do here use this data to look at the quality that data represents and also uh basically uh to see whether different data sets gives us different story in terms of mobility patterns in space and time throughout the pandemic in 2020 we started collecting a lot of data set from different companies and look at the patterns so here what i'm showing you uh is just four uh sources of data obtained from place iq the scarlet slab cubic and safe graph because these were the four sources that uh basically provided their user base number of cell phones scenes that were seen in their data set so if you correlate this with the population density of these data sets well the population density of the counties because these data sets are obtained at a county level um you can see some uh some of the counties that represent large metro areas such as Los Angeles also Harrison County in Texas Texas or um Cook County in Illinois there is an under representation of actual population with this data set so we need to be mindful about what data source we use and how the population is represented in this data set next we looked at about 26 different index uh indices from nine different sources and looked at the coverage in the spatial the spatial and temporal coverage of the data so how the missingness of the data or how much a coverage is available so here the color basically represents when the data is complete you see darker colors when the data is incomplete meaning that there are missing dates that are represented in each county you see lighter colors so um some of these sources provide really complete a complete representation to look at the year however some uh free sources like google visits and uh apple data sets there are a lot of um missing data in the data set similarly for the temporal representation of the data we have counties that we don't have data for several weeks so we need to be careful about when we are averaging and using this data set uh when there is missing information so then we wanted to explore uh and compare the patterns using different data sets and for that we use spatial autocorrelation and try to cluster to find clusters of mobility patterns throughout the COVID-19 pandemic at different locations using county level data that they're obtained at the weekly basis or average at the weekly basis so using lisa we created uh visual analytic tools that has ability for the users to explore these patterns computed with lisa and the basically patterns show hot spots and cold spots hot spots represent higher values clustered together with neighboring higher values and cold spots represent lower values close to other lower values so in general without giving much information at this stage because i'm going to focus on different states later and give you more information we can see a general agreement between cubic as a data of this data source that you can buy uh in this data so you need to pay for a license fee and save graph as a free source of uh data set so um and i'm showing two different indices one is mobility index or a median distance travel which basically represent how far people moved and then the shelter in place basically represent the percentage of people that stayed at home during this timeline so if we looked at these different maps we see a general agreement between the shelter in place data set between cubic and safe graph which uh there are a lot of cold spots in the midwest and south meaning that less people shared that in place and more people moved and there are some hotter spots in the northwest and east sorry northeast and west representing less people moved and more people shared in place however we don't see this agreement in the data source that represent the the distances travel and one reason could be because of the way these uh indices are calculated the algorithms to compute distance is more complicated than just counting the number of people so the tool is available uh at this link that i provided if you want to download and explore and there is an interactive version of it that i'm going to show next uh online so the visual analysis that we created has two components the first component is a hotter spot um and called a spot recency and consistency map which represents basically how often and how recent different counties were classified as a significant hotter spot and cold spot you can zoom in and see different counties in different states and basically the size of the dots represent the consistency in the behavior of being a hotter spot or colder spot the size of the larger dots represent the the frequent um cluster and the color the lighter color represents the older clusters in the timeline and the darker color represents more recent hotter spots and cold spots and then the next component is when um so previous component you would could see more spatial patterns here we are looking at patterns over time how consistent these clusters were over time so if you say an angled um line at 45 angle you you will see a county basically represent the county that was consistently a hotter spot or colder spot and horizontal lines indicates that uh the county changed behavior to not being a significant colder spot or hotter spots next i will um um basically uh this um uh representation indicated we can use that to look at how different time of timelines of covid impacted mobility patterns for example we can see in california and california people started to shelter more in place like less people moved as uh when they stayed at home order started but then in california we see a changing behavior that is when the uh show social distancing um maybe orders left that there was no significant hot spot colder spot and towards the end of the year they again there was an another policy which made people stay at home more so now here i'm zooming into georgia as an example to look at how the patterns show different um uh different information uh in terms of spatial patterns and temporal patterns so looking at these two uh uh data obtained from cubing and safecraft basically in general we see the same patterns differentiating ruler versus errant counties especially for atlanta which exhibit atlanta exhibits uh basically a hotter spot behavior but we see some differences between the two the data said although they both agree that atlanta was a significant hotter spot meaning that more people stayed at home however other places maybe lost um that are not in large metro area we see a significant colder spot that oftentimes also reacting to different timelines of covid or political situation for example uh we see a uptake of people moving more after the presidential election or uh around the timesgiving um like timeline uh holidays and also we see a change uh and generating more hotter spots around the second peak of the covid um spread uh around july 28th so another uh very interesting pattern uh is observed in south cacoda although we don't see that distinct uh differentiation between ruler and large urban areas we see a very distinct geographic patterns that is separated using based on meso sapriware so we see an east river region and west river region the east river region is mainly it makes mostly corn and meat agricultural production areas and west river region predominantly are ranch and uh dry land farming and mining operations and there are some uh indian counties that represent um native american uh majority in uh basically the west um areas so if you look at these patterns we mainly see uh hotter spots in the west and colder spot in the east however the recency and consistency of these patterns changes with the different data sets we see also uh we we don't see a very significant response to uh like all through time if you look at the timeline however um but and maybe it's because south cacoda didn't have any state wise uh stay at home or their situation here so um to basically recap what i showed uh here um there is a summary of all the states that we just talked about and the percentage of number of colder spots and hotter spots created based on the data sets used using travel distance and shelter at home so if we had consistency between the across different data set we would see all these colder spots and hotter stuff hotter spots would lie now however we don't see that here so i would like to basically summarize um these points on two different aspects of what i just present one would be focusing on visual analytics and how we can basically how we leverage visualization to inform our quantitative analysis so we basically combine qualitative assessment of what we see in terms of patterns with quantitative analysis of these patterns and what the data represents so it is important to bring the mapping and a qualitative analysis into quantitative modeling as well and also it is important to look at how this basically um visualization actually inform us so we need to do some user studies to make sure that these visualizations are not misinforming us either another aspect that i would like to touch on is data quality and uncertainty that we had a discussion about so basically one important aspect one issue that might come up is because these um there is no standard on how these mobility indices are calculated different companies have different algorithms sometimes the algorithms changes through time as well so even the comparison between the same index from the same provider might be fuzzy as well there is obviously a short date there is the bias in the data and representativeness issue that we need to consider and also these data we are not sure what population they are covering because we have these data so compute these data at like aggregate level and not a specific population level so we need to perhaps inform this data with more information about socio-demographic and geographic structures of these communities for example here if you look at the timeline of uh from 2019 which is before covid like the year that was of normal lay year and then the next year 2020 when covid happened you see a change in the behavior when higher income population started to moving less than before and where uh lower income population were more impacted and their behavior fully meaning that they even move more than people with uh of higher income and with that i would like to stop here and happy to be to answer any question and i would like also to thank bias students who contributed to this research thank you so much samaya that was really interesting and i just really appreciate um a critical view and a comparison of different sources of data that is so widely used and adopted especially as we saw during the covid pandemic so just wonderful presentation thank you so i'm going yeah i'm going to uh introduce our next uh talk which is actually by a team of speakers you have kate hess and matt woodleaf who are from esri or esri who will be presenting um uh in the next round so matt and kate i wouldn't there we are thank you so much we'll turn it over yeah thank you uh it's one of the thank you uh to the um the group for inviting us here it's really exciting when we get to you share some of the stuff that we've been been working on uh so we're going to go through uh a little presentation here and on using mixed methods approaches to understanding migration with the gis so since it's founding in 1951 the united nation's high council on rubies the unhcr has been tabulating displaced persons uh in 2017 the unhcr produced the story about the uc on your screen right now and at that time more than 50 million people have been forced from their homes due to war sectarian violence uh a natural disaster so if we fast forward just five years the unhcr now estimates that there are 80 well 89 89 million people have been forcibly displaced we felt that these rod numbers however we're not really telling the whole story we wanted to find a broader story by bringing together the different patterns of data and creating a narrative that could answer questions that says where are people fleeing from where are they going and are there countries that are more desirable for those seeking asylum and lastly are we able to tie specific government policies to an increased number of displaced persons so once we had these patterns of data at hand we wanted to get into the personal aspect of the crisis so in addition to looking at the refugee numbers we needed to assess refugees at the individual level and do some qualitative analysis so once migrating do the refugees feel safe are the conditions in refugee camps adequate how do individuals feel about the future and maybe most importantly what can we do to help so the process we followed was to quantify the migration of displaced persons first we wanted to know which countries they were leaving and which countries they were turning to for asylum to do this quantitative analysis we used a flow or a link map and we plotted the country of origin and the current country of residency by first spatially enabling the unhcr data set by geocoding the countries based on their central then we calculated the total number of persons of concern and symbolize the line to connect the origin and the destination immediately we saw the scope of the crisis there is not appeared to be a single country not affected by the refugee crisis and through this simple mapping method we were able to get some key insights first by symbolizing the country's country's circles by in degree we were able to identify countries that had refugees coming from many more locations we normalized the node values here to show the comparative difference between values of this person per destination country as you can see on the map the united states canada australia down here and norway and sweden make up a large taking refugees from all over the world so their nodes are comparatively larger than the other nodes on the screen we are also able to illustrate the flow of refugees by varying line thickness based on the calculated total so let's look at the syrian era of republics for an example we wanted to know in this case the line thickness here varies depending on how many people are connecting the origin to the destination so first you see syrians fleeing to turkey we see them fleeing to neighboring lebanon at one point we actually even see irakis fleeing into syria right but the most important or the most intriguing era to us was the one that is circling back in upon itself this era represents the internally displaced people is the thickest line on our on our chart our map so we can assume that most syrians did actually not did not flee to neighboring countries but sought refuge within their home so the map provided some key insights but as i mentioned earlier the unhcr has been collecting data on refugees since 1951 and the origin of refugees since 1960 i want to show you what the raw data table looks like in this raw data table we have nearly 120 000 records representing millions of people right so as such while the map could provide key insights it was much too hard to identify patterns and and gain a significant understanding of the situation but it still pointed us in an interesting direction so we decided upon using an interactive timeline to drive the other visualization let's take a look at 2017 so as i click through this interactive timeline from 2017 to 2021 i want to pay or have you paid special attention to this summary total down here of internally displaced persons so like in 28 27 to 2018 we see a while the numbers all change we see much bigger jumps within this population including this 2.3 million up 2.3 million person jump between 2017 and 2018 and then again another 2.3 million person jump between 2020 to 2021 and so this insight here trigger another round of analysis for us we shifted our focus to analyze internally displaced persons we wanted to gain an understanding if if or how policies established by government had any influence over whether people left the country entirely or stayed within the borders of their homeland and for this we looked into the Rohingya crisis in my mr as a case study so the Rohingya have long been targets of violence both by individual groups and the government and in fact in 1982 the citizens of the law specifically left the Rohingya off the list of national races and prevented them from ever getting citizenship this is significant because the Rohingya are now classified as stateless persons i'm starting in 2000s here in the mid 2000s the government began a household registration campaign in which families of Rohingya were gathered and a photo was taken and anyone not in that photo would not be part of the family and could not legally stay in Myanmar and as you can see in the bubble chart here many were able to find refuge in neighboring Thailand however Thailand soon enacted policies to slow the admittance of Rohingya into Thailand and we see that reflected in our data set in fact many of those policies began in 2006 and we start to see a couple interesting patterns emerged here first Bangladesh it makes its first appearance as a destination and we also start seeing our first internally displaced persons let's fast forward to 2010 in 2010 the government of Myanmar began its violent crackdown on the Rohingya many were now able to flee to Bangladesh and it became the country at which most thought asylum however in 2012 the Bangladeshi government officially tried to close its borders to Rohingya and in this case we see a decrease in country asylum being Bangladesh and a massive increase of internally displaced persons in fact it's 62,000 in 2010 it becomes over 430,000 in 2012 there's also in 2012 where authorities in Myanmar forced the relocation of Rohingya into impermanent camps and severely respected their access to travel work education and healthcare the last part I want to go look at is in 2017 in 2017 in response to an attack on police posts increased by Rohingya militants the military of Myanmar began a massive scale clearance operation and left with the choice of being killed or being forced to flee their homeland many Rohingya decided to flee to Bangladesh and we see that reflected in this part nearly one million people left the homeland to seek refuge and to this day in 2021 Bangladesh remains the country that most seek asylum into but however Bangladesh is not where our story is instead it's actually the beginning of the next phase of our analysis I'll turn it over to my colleague Kate great thanks Matt so we've spent quite a bit of time quantifying the refugee crisis but the crisis is more than just the numbers in the workbook we wanted to explore the human condition and really qualify the crisis so what do they need and do those needs have seasonality how do they feel about their situation and again how can we help so using geospatial solutions we're able to take the static data tables and tedious data tabulation and turn it into more real-time analysis this dashboard shows how survey results can be tabulated in real time the first tab here is the survey as it was received from UNHCR upon download it asks good questions but we wanted to add some additional questions that would better assess the spatial and temporal resolution of the crisis so we've added questions like how long have they spent in the camp and from which region of the country did they come we also added some semi-structured questions that would be more amenable to a qualitative methodology these allow us to use thematic coding and are less constrained than the other question types this is important because we're seeing gaps in the data versus the lived experience of refugees for example when it comes to sanitation the results from the survey show that a vast majority of people are satisfied with the sanitation conditions in the refugee camps but by conducting proximity and walk-time analysis we can see a different story each one of the rectangles here are tents that were detected using geoai methodologies from drone imagery shared by the international organization for migration after analysis it was concluded that 37 percent of the population living in these refugee camps did not have adequate allocation of washroom resources so that's nearly 12 000 people eliciting these gaps in the data set directs us to evaluate the survey instrument and can lean to refinement for the next iteration of the survey this highlights how by using a mixed methods approach we can use quantitative data to understand where a more qualitative analysis may be needed from our approach we realize there is much more we need to learn about the individual respondents to the survey and discover why their experiences were not that of the majority we've summarized this work in a story map where we can provide the background and context to the analysis and we'll share out a link to this as well the second half of the project we've shown today the goal was to evaluate satisfaction with the quality of life within the Rohingya refugee camps so this ties specifically into sustainable development goal six which is available clean water and sanitation now if we're just looking at the UNHCR survey findings we get the impression that sanitation is satisfactory in the camps and people are happy with the conditions saying quotes like that the available hygiene is improving but when we bring in the additional data sources and approaches like quantifying buildings and washrooms from aerial imagery and including more open-ended survey questions for additional qualitative input we're able to get a fuller picture of those conditions on the ground and in this case identify where there's still a lack of sanitation facilities ultimately having these more comprehensive data-driven insights enables decision makers to help people in the ways that they need it most so here we have a list of some of the changes that have been made to Rohingya refugee camps as of June 2020 including constructing new latrines and bathing facilities and increasing the number of households that have access to clean water and soap and we've included a link to the full report here where you can see additional details on the actions being taken on the ground to improve sanitation and safe water access so in this exercise we identified maybe the three main conclusions so first insights across a diversity of visualizations are necessary to tell a cohesive story rather than just the diversity of the type of data collected this allowed us to find gaps in the traditional narrative around refugees people will stay where they feel safe and in the case of Rohingya they had no other option but to leave their homes due to government policies secondly the iterative process between quantitative and qualitative was essential to understanding if the numbers reflected the sentiment of the Rohingya the iterative process allows also allowed us to refine our survey instrument and get into the why behind the where and the what and finally collaboration between the methods and data sets along with the personal collaboration allows us to dig deeper into the crisis and go beyond the numbers this collaborative analytic approach allowed us to see the story more holistically and to find questions that could help us fill in the gaps between the data and the lived experience with that let's turn to the next presenter thank you thank you so much Kate and Matt that was really really wonderful I loved the visualizations and this your description of the iterative process that really brought in qualitative understanding of what life was like on the ground in these refugee camps just really is a great exemplar of how we can do mixed methods research and integrate sort of the lived experiences of people on the ground so I'm sure you'll I'm sure the audience will have tons of questions but I thank you for your presentation and I would like to introduce our final speaker Marie Urban from Oak Ridge National Labs Marie over to you hi thank you let me share my screen oops I have to hit the share button sorry about that there we go all right looks good great thank you so thank you I appreciate the opportunity to talk to you all today and I've really enjoyed this whole discussion this afternoon as well so I'll be discussing the application of qualitative and quantitative data around IDP and refugees for population modeling so since the late 70s or NL has been involved in developing spatially refined estimates for populations at risk so initially in support of facility siding or shipment of hazardous materials and because of this research or NL develop landscape a global population data set for consequence assessment delivered in 1998 and continuous annual updates since 2000 it is the first global high resolution population distribution data set representing an unworned population so this unworn population is a more realistic spatial representation of population across activity spaces such as home work or school since the release of this global data set was nearly 25 years ago which by the way was well before geospatial data or the high resolution imagery that is now available and through leveraging our high performance computing computer vision and machine learning capabilities there is a continued refinement of the spatial fidelity of land scan with each new release our next release is land scan 2021 which will occur next month so this is a high level representation of the land scan remote sensing based global data modeling and mapping method I know cascade talked earlier about grid of population data sets and so that's kind of a nice segue into what I'm talking about today specifically about land scan this is land scan is a top-down multivariable asymmetric population modeling approach conducted by disaggregating census counts within subnational administrative boundaries so the land scan population distribution models are tailored to match the data conditions and geographic nature of each individual country and region so basically we develop a likelihood surface so where people are most likely to be found and this is performed using ancillary spatial data and high resolution imagery exploitation we then distribute the census populations to small areas within the census units based on the values in the likelihood surface the displaced population IDPs and refugees are then accounted for through data driven mixed methods so you know I don't have enough time to discuss everything about the land scan modeling but you can find more information out at landscan.orinal.gov so what causes population displacement and resettlement you know there's a lot of different events you know short-term events like conflict or persecution social and political instability natural disasters and then there's the long-term events right food insecurity water stress but then the short-term events can become the long-term events as well and then for you know kind of the typical background migration your census captures those movements so here the Kosovo refugee crisis was the first land scan update to account for IDP and refugee movements we used open source information humanitarian reports and media about locations and displaced populations to update the population distribution so by the end of April 1999 about 600,000 residents of Kosovo had become refugees another 400,000 were displaced inside Kosovo overall there were two million residents in Kosovo nearly half of the residents became refugees or internal displaced people so you can see that largest spike in the graphic below is about 375,000 Kosovars move moving south to neighboring Albania and then there's about 150,000 who moved to Macedonia and you can see those spikes throughout the graphic below and the movement of the refugees and IDPs so normally we identify new or monitor continuing events of population displacement through humanitarian or other open source resources however you know you know some events are rapidly evolving situations that require assistance immediately so if we use Ukraine as an example being the most recent continuing conflict where bordering countries and humanitarian organizations provide housing and supplies to IDPs and refugees the other example is as kind of piggybacking on Matt's discussion about Myanmar the Rohingya how in August 2017 there are nearly one million Rohingyas that refugees fleeing to Bangladesh to settle in Cox Bazaar refugee camp and then of course smaller numbers of refugees in surrounding countries for these types of situations and this has been discussed all day there there's often a shorter term need for information that can aid in estimating damage and providing relief to effective people so there isn't you know a single authoritative reporting source but rather multiple organizations and media that report the rapidly changing situations for example UNHCR which we we heard about earlier global conflict tracker you know displacement tracking matrix by IOM and you know there's even with all those that reporting there's less reporting on the resettlement of refugees especially IDPs so it's often difficult to track the resettlement and last of all aligning disparate data sources and earth observation data for population modeling is it can be very challenging and so what happens oops sorry my apologies is that we end up with a really complex challenge and trying to account for refugees and IDPs in our population modeling so for population modeling on account of refugees you know from the qualitative data side of things you know there's cultural practices of things like offers to refugees or IDPs to stay with local families and so those are hard to track but sometimes you can get information from anecdotal reporting maybe from some of the discussion within media and other organizations about displacements like kind of the location to and from there's often behavioral surveys you know they they give some insight into where refugees IDPs were maybe initially displaced and then re-displaced we can also we have the the capability to detect change in growth and decline of camps and so that change can lead to pursuit of new sources to understand the population growth or decline of those camps and then most important for population modeling is understanding the building use and so if there's any changes in that building use it's imperative to understand that we often see building use change once there's a displacement event Ukraine for example people use museums theaters and churches as shelters which then changes the building function and the expected number of people within the building and then for population modeling that's significant the the last thing is you know it's useful to identify the destroyed infrastructure you know if a building is destroyed we consider it inhabitable and no populations would be distributed to that location so for quantitative sources we regularly collect open source and media population estimates to inform our population models for an area over the longer term humanitarian organizations and NGOs report registered camp populations those can include detailed camp descriptions and demographics so then we've also developed tools that exploit imagery and support of developing camp population estimates or capacities when those aren't recorded so here's an example of using overhead imagery to detect camp establishment and growth these images show the rapid construction of IDP camps due to the mass muzzle exodus between October and November of 2016 so using counting tools we can rapidly determine whether the capacity of camps supports the reported displaced population in addition to comparing with reported counts we can then pursue where the unaccounted refugees or IDPs may be found so this is also an example of how alignment of a reported population estimates in EO earth observation is important for land scan updates here's another example of using overhead imagery and reported estimates to understand population movements so border crossings are important for monitoring population movements and identifying camps this example is the Syria conflict and the opening of a former closed border into Iraq here the number of refugees crossing went from 68,000 to over 200,000 over the course of the year so what we typically see is that most refugees settled just past the border for minimum displacement but also in hopes of return so we pull information from social media as well for most more insight into refugee and IDP activities so these photos were pulled from discussion about humanitarian support this is an example of two different impromptu facility uses that result in differing population estimates over time so the facility on the left is a shelter and hosts a large population whereas the facility on the right is a distribution center so there's a large gathering of people outside the facility for the distribution center but that is only for a short time a short period of time for that day this example this is an example of our capabilities in automated change detection to determine habitable buildings so leveraging high-performance computing and machine learning capabilities so our algorithms are quickly detect buildings prior and post-conflict this is a maria pool so the damage buildings are rapidly identified to determine displaced populations and inhabitable buildings for land scan updates populations are not distributed to the destroyed buildings and so that's key for us to understand those here's another example of rapidly detecting change through building detection and counting to get us closer to the number of people affected by conflict in this case it's the Boko Haram who terrorizes city resulting in the loss of infrastructure and life so using our building detection in order in addition to accounting capability the number of sorry the number of structures remaining quickly provides insight into how many households displaced sorry let me go back so we have over 10 000 structures identified here and counted and then under 10 000 after the Boko Haram moved through the area here's a close-up where you can see the structures and the count and then afterwards really see the damaged or destroyed buildings and and your account it should be noted that this example occurred during a time when you know digital trace data wasn't readily available and it may not be available yet in many areas of the world and so being able to monitor see where people are moving isn't you know if we're looking at the global scale quite available yet however Ukraine was a great example of that happening I've seen many demonstrations of watching the the populations moving outside of Ukraine to neighboring countries so Orinel developed the capability to continuously identify areas that may be experiencing power outage using beers nighttime lights this is a dashboard the the legend on the left shows black as decreased lights white as no change yellow as increased lights and gray as cloud cover so in this image there's quite a bit of cloud cover but you can still get a glimpse of decreased and increased lights in parts of the country this capability supports rapidly detecting changes in outages especially during natural disasters and gives insight into who is affected and may need support so if communities are without power for too long they potentially begin moving where they can find power and shelter so last of all this graphic is a visualization of population dynamics during the early stages of the Syrian conflict so by now and the estimate of internally displaced people is risen to between one and a half and two million the intense fighting is in densely populated urban areas cause more and more sorry the the intense fighting has caused more and more people to flee their homes seeking refuge in schools dormitories mosques and other public buildings so many have been displaced more than once fleeing from hams to Damascus for example and then moving elsewhere in search of safety and so you can see the large movements of population since we did a a subtraction of the population to get a better understanding of the movement so understanding the losses and gains you can see many of the populations at this time and this is of course in august 2012 have moved north outside of syria and south as well and that's the end I'm happy to stop sharing thank you so much Marie again another wonderful example of how we can use sometimes called nontraditional data sources but in reality lots of different types of data that have both a qualitative and a quantitative aspect to it to help us understand really complex social problems so thank you so much for your presentation I first would like to offer the opportunity for the committee members to ask a question if anybody on the committee go ahead and raise your hand in zoom if you'd like to ask a question and I would say again remind public attendees watching the webcast that if you have an idea or a question that you would like please put it in the slido and we'll work those into our conversation at the end of this panel so committee sorry any questions harvey please again thanks for thanks for another great panel I don't know if this is a well-formed question but something I've been thinking about is that when we talk about mixed methods it seems like we're still loosely coupling quantitative and qualitative data we can see that with story maps we're basically we're telling interesting narratives but we're just basically georeferencing like media I'm wondering if you could speculate what are better ways to think about next generation of integrating qualitative and quantitative information I mean is is there a better way than just you know um measurements and exploration or using qualitative data to to inform how we do our quantity how we interpret quantitative findings is there a better way to bring these data together like I said that may not be a very well-formed question but I I just have a feeling that we're still in the first generation at least in the geospatial realm of integrating these types of data and there there are more advanced ways of doing this yeah I guess I can take a stab at answering that question um so I guess maybe one of the more um exciting things coming out of geospatial is the ability to map natural language processing so um from for example the the gdel that we saw earlier um we can take that information and put on a map right which is half the battle finding out about anything so that the context really does matter of where things are happening um so I think going forward and seeing those future developments in the natural language processing and being able to get the data faster rather than waiting or reports to come out or being able to download some data and have to explore to find our narrative like Kate and I did um I'm optimistic that can be something in the future that will help us explore data faster and understate it faster maybe provide a quicker link between the qualitative and the quantitative I don't know if that was a well-formed answer but that was my it's as good as my question thanks Matt are there other panelists Maya yeah I I would add maybe um you know using leveraging visualization and visual analytics maybe we can develop more flexible visual analytics tools that are better coupled with like both bringing both qualitative and quantitative data set together so in the background in visual analytics we have data machine learning techniques that are totally on quantitative data using quantitative data but then we map the results can we do that on the fly by integrating like information from imagery and as Matt said like natural language processing to mine a qualitative data as well as quantitative data and map it that would be interesting thank you so yeah so um Marie really quick um I'll comment I think one thing that we're experiencing and I I think it was discussed earlier is that we're in this huge you know we're we're drowning in data right we just have so much of it and I think you know at this point we really have to focus what it is that we want to to answer or we'll get out of of the data and we're we're in a position now where it will we can't just I don't know I don't know if anyone does this anymore but you know use one or two you know data sets it's going to be pulling in multiple so you know like I talked about earlier we we pull in so much different data and it's almost you know making sure that we have if it's qualitative quantitative multiple sources to to verify you know what's the quality of this data as well and so you know moving forward I mean I think we'll continue to develop amazing tools to sort and sift through and understand the data and be able to answer you know continue to answer some challenging research questions but also um hopefully take into account um you know how how certain the data is as well hopefully sorry if that was a little crazy but a little different oh that was good thank you thank you all yeah thank you so much um I have two hands Riz Riz and I think uh kirsten you were you had your hand up first yeah thank you um you you sort of answered some of the questions that I had um and harv um was touching upon them because it seems as though your session really wrapped up nicely what we started in the beginning of the day where we were talking about um you know the real-time data and Anna was talking about the you know the need of that and then Lydia got into more of a design process and then you know Steve was looking at this communication um and so I guess it's sort of a question about again back to all this data and Marie you in particular I think we're starting to bring it into what are the crises and what data is needed in each type of crisis that's happening and it seems as though all of these um approaches need different types of data and so I guess maybe a question is is anybody out there trying to organize this into here's the data that you need you know for this type of crisis because there is so much data out there flying around and I think it's it's uh almost overwhelming for organizations and policymakers to try and figure out what bucket does this fall into so is anybody doing any kind of work related to that and I don't even know if that's even a question or a comment or just thoughts about it yeah I mean so I've uh I remember as it pertains to the Ukraine crisis some of the departments in in the UN and WHO uh have been trying to build collaborative environments where they can transfer data back and forth but still maintain the integrity of their own data sets so they can still choose what to share uh maintain that control over their own data but make sure that the data transfer is faster so rather than having to to wait right it's that's a big problem we have to wait for this data well if you have it I know that you have it we have a collaboration let's just spare it um right away right sorry and these are acting up acting upon it so I think we'll see so maybe more of those really more of those collaborative environments to to share data uh faster could be could be a key in the future I can I might say can I can I add one thing so one of the things that I've observed is that there's you know while we have a lot of standards sort of metadata standards that have been created for quantitative data um and and there's there's not one metadata standard although there's a lot of effort I think that's been put into trying to create standards around how to structure data there's still a lot of sort of diverse um you know different groups are trying to develop their own so I work with the food FAO right and that world food program and they have just created this new dataset called gift which is supposed to curate and bring in all these food intake and food related datasets from around the world and the metadata that they have on their website is structured differently than metadata that I would pull down from say the iphone's effort which is looking at more population surveys so there's still this sort of diversity in what's required of a dataset to kind of create a repository or even information in an informational repository about what's out there I would say one of the great challenges in qualitative data is that those standards really don't exist right and it's partially because of where this is coming from and what disciplines it's coming out of the more qualitative side of the research that's been conducted but I also just think there hasn't been sort of more of a top-down like we've got to create some standards for qualitative data so that we could more widely disseminate it to people who might be interested in integrating it and using it so I'd say that's actually a gap um that that we could um look into and and perhaps fix although if not everything's fixable but at least start to structure that data so it could be integrated and more widely used um Samaya sorry yeah I also wanted to touch on that in regarding a metadata standards and having a uniform policy or like a centralized way of like collecting a repository of mobility data that can be used for looking at you know how mobility is changing in response to disruptive events so the problem with mobility data is that they are collected by commercial companies and some organizations and there are different policies different standards the different ways of computing and movement indicators and the problem is that um a lot of these companies sell their data right so these are not free data sets and like for example like satellite imagery is land news data these are easy to be coordinated with different companies and have a centralized repository or a data source for that however with mobility data is basically the at the discretion of the data owner how they want to share it how much they want to share and how they are computing it and there is no effort to bring all these different sources of data together and that would be actually a challenge and next step to look into how we can create a standard or central repository of such data sets excellent yeah thank you so much I think this sort of idea of standards around qualitative data is one that is really important actually as there's sort of uh you know we talk about the exponential growth of data and I would say that that's no different for qualitative forms of information that we could use uh as data in our in our research in our analyses Miguel I see you have your hand up as well yes thanks Elizabeth I am very curious to hear perhaps some Kate and Matthew's opinions about sort of the proliferation of the digital twin model in in this area and I saw hey Matthew I saw you grin so you know we're all geographers I think in this room and we've been I mean my first SRE conference user conference was in Taiwan was in 2001 and I was in a meeting like 21 years later in Europe and they're talking about these amazing DRG and it's like that looks a lot like a GIS and are we repackaging you know I think they were trying to repackage a solution that is based on very fundamental geospatial analysis principles and it's partly I don't know maybe it's because industry is not aware of the state of the science what's the disconnect and what's sort of your opinion and there's a do we have do I have to now every time I wrote a proposal replace the name GIS with zero twin to get it selected I mean like I don't I don't understand what's going on so I just want to get your perspective within this context because I think there's sort of a integration gets into the issue of the integration that I spoke about earlier is that the solution or do you think what are your opinions personally you don't have to talk about you know your code you know your company's perspective but yeah I can take that one so so I would say with digital twins it is a bit of a just kind of fashionable term right now but I think it can kind of incorporate existing GIS concepts and then in some cases extends them a little bit further so I think just having digital twins exist at such a different range of scales you could have a digital twin of an entire country or of just the process is happening in one factory and I think digital twins really need GIS kind of as a basis because they're capturing the relationships between different processes and objects and that inherently is going to include how they're related in space so I think it's adding something a little bit new maybe to the existing field of GIS and what has been already kind of the standards of how people work with data but I think it is adding some new benefits to kind of how we can think about data and how things are related to each other and gives us really that ability to take things a step further from just visualizing how things are interacting now and performing kind of projections and looking at different future scenarios and how that will impact all of the different current components of our system of our digital twin and just wanted to add that I think the digital twin concept is really is about that data integration GIS is just a tool to help integrate that data but what you do with it at the end can help you build out so-called digital twins so what systems are you modeling the interplay between the natural systems in the built environment um how everything interacts and then how you can run analytics and models on those I think that's what kind of turns the digital twin from like a 2D GIS type of thing to a 3D full digital twin I guess I guess another aspect of it is the scale of digital twin like how a granule you want to have that model and how that invade the privacy of for example if you think about mobility I think that would be an issue to consider thank you I do want to allow one question from the from the webcast audience and I would like to bring that into our discussion so I actually have a very specific question about diastometric mapping and so the question was have diastometric mapping techniques been validated specifically with refugees and transient populations and the more specific question is do you use the binary method or something else so if one of our panelists has used diastometric mapping has a thought on that it'd be great to share um I think I can start I think um so the first question sorry Elizabeth is um has it been validated or was that specifically yeah validated specifically for refugee and transient transient populations um so I you know diastometric modeling is is really getting at how can you use you know how can you what how can you take this census data and disaggregate to what you know and and you're identifying if it's if it's a building um what type of building that is and then as you're talking about refugees and transient populations um those um fold in nicely because you're saying okay if we have a refugee for example if we have a refugee camp here's um how many people within this area um as I spoke before about you know pulling in different types of data to as as kind of um indicators of where people would be and um for the transient um you know and validation you know validation is hard um first of all because you know our if you were to like just look around you how many people are here we can count how many people are here at this moment like I can say I'm in my office at this moment how many people are in the building but in another minute that changes right and so validation of the population of those estimates is difficult from that perspective and when you start throwing transient populations in there that adds another layer or a dimension of um of difficulty um then your account you know how are people moving and um running simulations and understanding that movement and those patterns helps and can help with that validation um but again we get back to you know no one's staying in one place for you know um an inordinate amount of time and so so that that that makes that validation part uh a difficult thank you any other plots from our panel on dice metric mapping so we only actually have two minutes left um in our panel this afternoon um and so I did want to ask one quick question that we had discussed as a panel um and I I think that I think that one of the really interesting things about the different approaches that you've taken is that um you examine the dynamics of migration and displacement at very different spatial scale right and we're all geographers and so the space time nexus is we know is very challenging um but I might ask quickly for like two sentences or maybe maybe four sentences on how if the temporal scale of analysis changed the question uh how would how would the question you're asking your data change right so like yeah anyway so I'll leave it that maybe I'll I'll I'll go to Samaya first and then we can popcorn sure uh so um we were studying uh COVID as a you know even that we were looking at and how that relates to mobility so for that um the temporal scales that we need to be able to relate patterns of mobility to COVID was at a daily or weekly basis so if you had the data that comes in like monthly resolution then that would not be valuable valuable for the processes that you are trying to explore so I guess that it is important to understand what processes we are looking at in terms of spatial and temporal scales and whether the data actually matches and there's also temporal scale of the analysis matters that processes that we are trying to capture um I can um so for population modeling we're always moving towards a higher temporal aspect of understanding where people are and um I think you know Samaya's work she's you know getting at that you know minute by minute or you know high resolutions uh temporal scale um which is um pretty incredible and and um given where we are or what I've been talking about today and trying to capture uh refugees and IDPs you know just understanding um those at um you know maybe over a a different temporal scale which is more of months to years given where they're at and how they're moving um it's kind of puts it in a different dimension if that makes sense um and and changes the that question of temporal scale for our presentation I'll let Matt kind of talk about his section that was looking at the multi-year analysis but I think it's interesting with the survey and some of the drone imagery analysis that the section we were looking at today was really kind of a snapshot in time of just surveying people and how they're doing on the day they're surveyed and looking at the one day that the drone imagery was collected but I think it can get really interesting if you're collecting the information multiple times and then can start to do a comparison of how things are changing and maybe evaluating how the if you're looking at drone imagery that's taken on another day and doing a change comparison to just identify how many new facilities have been built and recalculating that distribution of facilities per per person and then comparing that with the more qualitative survey results and seeing how closely tied together those are with people's changing perspectives of the conditions versus what we can actually see or the conditions from looking at imagery. And looking at our temporal scale of the quantitative analysis um as unfortunately it's only it's only a year is year by year so I think it's the aggregation of that year but something that we were interested in exploring was the seasonality of migration. Do you people move more often when whatever weather is appropriate for it right? Does weather have an impact of it? Does it have any sort of relationship between what we call the fighting season and the off season? Do you people move sooner after you know holidays or religious events? Just trying to be able to find that information would have been I think would have been interesting to see how that changed the flow patterns of migration to have more granularity in the in the seasons or the or even months per year. So we're looking at our insights looks just like all right in one year all these people moved but they didn't move all at once right so people like to go back and try to understand now are they moving what's in the country are they especially the internally displaced persons are they coming from one spot to another or they're making multiple stops along the way and that wasn't information that we had that would be really great if you could find that. That's always the trick isn't it? So with that I believe that ends the third panel of our of our sessions this afternoon so I'd like to hand it back over to I believe Harvey to close us out for the afternoon. I thank you all again speakers it was a wonderful panel I really appreciate the time and effort you put in to sharing your work with us. Thank you. Yeah thanks for having us. Thank you. Yeah what a great afternoon thanks to everyone thanks to panelists thanks to our keynote speaker and thanks to members of the MSC and GSC for moderating these sessions lots of food for thought we'll be discussing this internally and thinking about what are the next steps that we can act upon all these wonderful ideas and thoughts and challenges we heard today. So on behalf of Pat McDowell the chair of the Geographical Sciences Committee and myself the chair of the Mapping Science Committee we're calling this meeting to a close I want to thank all again all the participants but also the audience for listening in I hope you learned a lot today and I can go forward and keep pushing the envelope in this very important area. Thank you all and with that we'll sign off.