 We are live now. You can start in two minutes. Thank you. Okay, that's a lot to come on see me Who is the same as you start Okay, well good morning. Good afternoon or good evening everyone. I think we will go ahead and get started with the webinar My name is Scott Ross Patrick. I'm the DHS to analytics product manager And I have the pleasure of hosting this webinar this first one here is in English and then we will have the exact same webinar In french immediately following this So you are here to hear about the various dhi's to community innovations for vaccine delivery and micro planning We have three of these Innovations lined up to speak to you today really exciting stuff really kind of cutting edge approaches and technology methodologies Micro planning and vaccine delivery So it's going to be really cool to see these folks present Let's just go ahead and Take a look quick look at the agenda So of course, I'm going to give you a quick introduction I'm going to provide a couple of additional information some perspectives on innovation within the dhi's to ecosystem specifically And then we have our good friend koreen at gavi going to come in and take us through some of the Principles around micro planning and vaccine distribution just to make sure that we're all on the same page And then we start with our innovations. So the first one we have is an organization grid three They have three presenters silvia ren hether and nyntow will all take us through their innovations around mapping population mapping Then we have another really cool presentation from flowminder Oil it is going to take us through that looking at population mobility Approaches the various tools that they have at their organization develop to support Vaccine delivery and then finally we have a video from rafael from maha And they'll be taking us through their Their applications and tools for micro planning as well um One quick point to make before we start to get into the actual presentations is that we have We request that you ask any questions that you have from this webinar on our community of practice And I believe that alice has already posted the community practice link in the chat So please look at the chat follow that link and then ask your questions in the community practice The reason that we do this is so that we have those questions In the community of practice and we can refer back to them, you know, essentially forever They are going to stay in the community practice. They're not going to be temporary like they would be in the zoom chat And that we can continue to answer questions. We can get more people to come in and And weigh in on the questions and get more perspectives and just have a dialogue in the community of practice Where we're really not able to do that in the um in the zoom chat So please do click on that link go to the community practice and ask your questions there If you have any or comments suggestions thoughts ideas insights all very welcome in the community practice, uh, of course All right, so let's start talking a little bit about innovation. What does innovation actually mean? From in the dhi's to ecosystem or when we're talking about kind of various additions to dhi's to Well, really we have a We have over the last couple of years had an absolute explosion of innovations third-party innovations in the dhi's to ecosystem and what I mean by that is People have been developing applications. They developed tools. They developed resources. They developed metadata packages They've have libraries Various data sets and These are done by third parties. These are not done by the university of oslo These are done by organizations that use dhi's to appreciate that they need additional functionality or features or resources or data or whatever in their dhi's to instance or in the dhi's to instances that they work with in order to support Some specific program or project And what we're going to zoom in on today is a lot of the incredible innovations that have been happening around vaccine delivery and micro planning But here you can just see a quick screenshot of you know Just a handful of the dozens of applications and innovations that have come up in the dhi's to ecosystem over the last You know a few years and some of these very very complex very cutting edge Other these are quite simple and but but effective Um, and and I think we'll hear a good selection of of of both of those today so When we're talking about innovation one thing that we have to appreciate is that dhi's to is a platform Right, so kind of just like you have android or your ios platform for yourself for your smartphone dhi's to is similar. It's application based and what that means is that third parties Can develop new applications develop new tools within dhi's to And then they can make these publicly available for any dhi's to user to be able to access And and employ in their own use cases A lot of times this means that these innovations are Filling gaps or addressing shortcomings that the core applications Um Have and of course we at university of oslo appreciate that core dhi's to applications and functionalities are not going to be able to cover the very broad Um degree of requirements that dhi's to has It crosses various countries and implementations and we try to build an ecosystem We try to build the resources the libraries the standards that allow for organizations like the ones that will be presenting to Today to build on top of what we have to address specific requirements to to fill those gaps Where the core doesn't um doesn't cover And what this does is this means that It's not just the university of oslo generating value or content Or functionality within the dhi's to ecosystem It's a whole world of contributors innovators that are feeding into this and we're all benefiting from it Of course here at the university. We're able to see these innovations We're able to share these innovations, of course more broadly, but we're able to study them and see You know, are some of these things that they have developed? Um Organizations like maha or grid three. Can we bring them into core? How do we make them more accessible more? Better performing more stable for everyone who might want to use them Um, and so a lot of times innovations that you're seeing within the core applications those kinds of functionalities that we highlight on our release notes Um, those are sometimes not new to dhi's too. In fact oftentimes they have been in dhi's too But just developed by a third party Uh, you know, they've been in a third party app and developed by An innovator that's not associated with the university of oslo But they have had such an impact that we have said no, let's bring those back into core Let's make sure that they're accessible or perform it and sustainable for everyone Um There are a few ways in which we kind of share these These innovations of course through the community practice. So if you're not a member of dhi's two dot org backslash community Um, or sorry community dot dhi's two dot org Please go into the community practice. Of course the link's been provided in the chat already We have an app hub. So if you're developing apps, um And you want to share those applications with all the dhi's to users around the world You can post your app into the app hub And some of these apps that you'll be seeing today are available in the app hub We of course have webinars like where we are today and we try to do as much training as as possible So we have our academy program that go through A lot of the dhi's two functionalities and we also have academies for folks who are trying to build applications So we have developer academies. So if if if you're kind of inspired by the presentations today You say, hey, I need to I can develop something in dhi's two Please do sign up for some of the developer academies and and we'll give you the resources and tools you need to get that process started All right so today we are going to take a little bit more of a specific focus on innovations for vaccination of course vaccination being a Critical thing that all of us are going through experiencing in our day-to-day lives But also many of us are supporting through covax as well and And of course dhi's two being used in a few dozen countries in covax response and these various vaccination innovations are really like I said quite cutting-edge they are Pushing the boundaries of of dhi's two in many regards. They are in most cases far beyond what we are providing in the core and And you know, I think that they are also broadly applicable Not just to the countries that they are supporting us as individual organizations or projects But broadly to all dhi's two users So even if you know, you have if you're not working with maha today or grid three or flowminder now in your country or projects I think that the innovations that they're going to be able to share with you Could be broadly applied. They could be available to you. They you know, they they they are They're generic. They're accessible and and I think that's the really exciting thing Is that they are developing these innovations not just for one country or one project all the time? They're thinking broadly. They're thinking globally And that means that everyone can access and access them So I think that's my last slide. That's just a quick introduction. I will now hand it over to I'll stop sharing hand it over to green to take us through the The goffey foundations for vaccination. So over to you green and you're muted sorry, so Let me try again and Hello to everyone. I'm very happy to be here and a big. Thank you for to the university of oslo for the invitation So very quickly, you know, because there's the most important Important presentation will come after I just wanted to to come in on behalf of gavi to say how this topic is Important for for us. I'm karen. I'm working at gavi the global alliance for vaccination And I'm the focal point for digital health information. We have a strong partnership between gavi and university of oslo and all our alliance Members are also contributing especially unicef and wh show so just to share that I think two years ago now we we really wanted to know what was happening in the gis landscape and actually we we really Went through a lot of review of existing projects and all the activities that we're going on in countries We support from many of you and we came up with a list of eight use cases Where we saw that, you know applying geospatial data and technology can really help to improve the immunization program And you will see that micro planning and digital micro planning is part of this so based on that Review and you know all the lessons learned we included digital micro planning gis transcending and all those elements Into our top priority for the digital health information strategies that we are developing Now so you can see that it's really coming within the first top priority, which is the to help us to Identify what we call the zero dose children children who never been to vaccinated in a health center And also to to identify them and then to reach them so I just wanted to really emphasize how Actually the the the micro planning especially digital micro planning can Help countries and and all our alliance efforts in this framework what we call the irma framework So is the identification of those zero dose children and the mist communities and then we can reach them and then of course we can monitor the the the progress and implementation of the of that micro plan So you see how important it is for us and i'm very glad to see many people connecting to the to learn more from that I just also say that you know Digital micro planning has been you know seen as very important for routine immunization But also in the case of the cobin 19 vaccine delivery And you will find here we have the our guidance on the for the funding on cds funding delivery for for cobin 19 vaccine And actually digital micro planning has been highlighted as one of the first Example of innovative intervention. We are encouraging and country can use us our funding to support that So quickly because you know coming just with great intervention without Mechanism to implement them is a bit frustrating, but we wanted to reassure you that We are trying our best to support countries And we have two ways the first way is with technical assistance and the second is for operational cost So in terms of technical assistance just to say that for the countries themselves and all the partners You know just spread the word for for countries who are eligible for for gavi funding There is an existing mechanism for technical assistance for the his2 We have a multi-country agreement between the gavi university of oslo and all the isp To support all those countries if you hear about any need You know people can contact me or the gavi senior country manager the isp of course If it's related to cobin 19 our support is broader it goes beyond just a gavi eligible country We can support the 92 IMC countries so the same is that you contact gavi or unicef for the 32 countries with direct support from unicef And the hisp as well in terms of technical assistance as you will see You know this topic is not just the his2 is also around gis we have put up and we are supporting different Mechanism one is the other dice with a digital health center for excellence with wh show unicef So you can click on the link you can see how to get support from there and then you know, we are supporting them There is also the great group the gis Excellent center and wh unicef are working hard on on supporting the country And you also see that the wh show unicef they have mapped all the gis technical partners. There are more than 40 That are part of that group So for all of that there is different mechanism not always easy to manipulate them But without trying you know to put that technical assistance network available to the country Beyond technical assistance. We know that the key of the You know, it's really funding for all the implementation the operational cost And we want to remind that you know for all the country if it's for routine immunization, you know, they can heavily use their hss grant new What we call the equity accelerator fund and also the campaign grant to support implementation And roll out of the hs2 and micro planning effort and to even go beyond we with the collaboration of unicef and a lot of experts who are present today and the health enable We provided the guidance and budgeting guidance on how to to budget what you need for jis transcending and then you can plug it plug those needs in your grants for the covin 19 Effort implementation. So this we have those funding available for the 92 countries You will have that presentation and all the links are embedded so you can navigate more easily in all those information We have also supported the development of that guidance from dice With the hs2 and jis and we've seen that guidance in same you click on the link You will have a lot of information on on the planning on the identification of target population for example And here's a link for the hs2 unicef gs working group and the list of all the technical provider I did not include a slide on how to get you know to the c.o.p. For the hs2 You you are already the c.o.p. I think during the webinar So really a big thank you for university of oslo and isp and all the partner including My hard-reaching reminders that who are partners and there is many others in the call So big. Thank you to everyone And don't hesitate to contact us at b.i.v. If we can help Great. Thanks so much kareen. All right now. I think that we get into the The reason that we're all here which is start to hear from the innovation innovators. So I think silvia you are Up first and sharing your screen. Is that correct? Yeah Can you see my screen? Yes, looks great. Great. Okay. Thanks. So hi everyone and thanks for giving grid three the opportunity to present um, so I'm presenting on the grid three project and on some of our data solutions for immunization interventions And some of these solutions can obviously also be used in dh is too and I'll say a little bit about that during the presentation as well grid three is funded by the bill and melinda gates foundation and fcdo formerly diffid We are a partnership between Season over at columbia university in new york the managing partner Then we have world pop flowminder and unfpa and we also collaborate Many times with frame Um And these are the presenters on behalf of grid three. Um, as I mentioned, my name's silvia ren. I'm the grid three strategic partnership Manager. Um, I worked in southern africa for a very long time Um seconded to governments. Um, I worked scott for for a bit as well Um, and also kind of with like, you know, I'm implementing dh is too in in various countries Um, we have two other presenters apart from myself on the call today and I'll let you heather and yank tau Introduce yourselves at this point. Thanks Hi, so I'm heather chamber. I'm enterprise fellow with a world pop at the university of south hampton and involved in the grid three project in a number of countries And um, so hi everyone. My name is in town. I'm a lead technical analyst at frame I have been involved in the cover risk mapping and vaccine prioritization work with three three since early 2020 Um, nice to meet you all So I'll give an overview of the project So basically our mission is to build spatial data solutions that make development goals available So spatial data is any data that has a location in it And I just wanted to walk you through some of our geospatial layers So we create spatial layers or for spatial data layers together with governments This kind of ensures the buy-in And also the use And I'll be walking through some of these spatial data layers now And so will heather and yank tau so that you kind of see what kind of data is available So one of our kind of flagship products is the settlement extents These are available for all of sub-Saharan africa And basically what happened here is you have a satellite who took pictures of all of sub-Saharan africa There was an automated process to extract all buildings And these were then kind of buffered into settlement extents And this is really the first time that a lot of these villages or settlements have been put on a map Um, and here you can kind of see the extent of this data set So all the countries you see popping up there, we have settlement extents for all those countries Um, I dropped a link below into uh, this this slide But I we can put it in the chat as well. So all of the data that we'll be showing you is also available for download and use Sorry So the the data I just showed you before was the location of the settlement kind of polygons as we call them They do not yet have names. We also support governments to add names to these settlements as you know Um, those of you who are working in in sub-Saharan africa sometimes settlements have more than one name Um, so so it's it's a bit more difficult and a bit more time consuming But we do collect those names And typically we use data that's already existing through census cartographies Or um, you know, or we sometimes also help collect those data sets And I just wanted to show you an example of how we've used those settlement extents in planning This is a very recent example from zambia We are we have been integrating these maps into the reach every district reach every child Microplanning template that is now used for cobit 19 outreach And here you can kind of see those settlements the settlement extents And you can also see some population numbers Heather will be talking about that later a little bit later on And the government for this to be in line with their template asked us To color code with those settlements according to distance from the the assigned health facility And so they are using this to you know to pinpoint hard to reach settlements and to make plans on how to reach them This is being rolled out all over zambia And every catchment is getting a map like that that is attached to that micro planning template We also help to compile or create other types of of data Such as health facility data sets a lot of time You'll be surprised but many countries actually do have health facility locations available But they might be in various different data sets Or they don't harmonize we see this a lot. We have maybe the coordinates, but they don't have the dh is to unique ids So we also help with that that type of matching so that you could get those coordinates Into the dh is to You know, and here's an example how we used it to plan a mobile vaccine sites in Ethiopia This time we're also using distance and using distance is a is a is a big advantage of the gis So as you get location-based data, you can you know model distances And here we have the distance based on travel time. This is walking um, and then you can see Settlements here that fall outside of these red zones Where people have to walk for more than an hour to get to a health facility And obviously the terrain as you can see here doesn't make that easier So this is a different type of modeling than we did for the Zambia settlements where it's purely based on distance Which some governments like because it's in their strategic plan Or also based on travel time. So there are multiple ways to kind of map that We obviously also help with administrative boundaries That can be the official admin boundaries matching a province to a district if there are maybe new Administrative boundaries that have been created to kind of help put those Into the gis and obviously those can then also be added to to dh is too But we also help with the model or the creation of of catchment areas um, and there are many different ways to do that um This is just one of the many ways um, so here we basically took population And this is another case from Zambia where basically we want one community health worker per 750 people And then you can draw these boundaries around the 750 people where you you could place a community health worker In the case of Zambia, we actually had the locations of existing community health workers So we could really nicely see areas That weren't yet covered by a community health worker and that can then obviously be used to plan community health workers And we also provide capacity strengthening. So when we create this data, it's not in a silo sitting somewhere We work with the government. It would have to be the the correct entity of the government We work with them to create this data So if we're looking at administrative boundaries a lot of times it's ministry of lands If we're looking at population data, it would be the stats department health facilities and the Ministry of Health So we co-create that data Then we also give, you know, just the general GIS technical capacity strengthening lessons in the classroom or online self-paced learning But then we also have modules for decision makers on how to use the data Because unfortunately a lot of times the technical people who are working with the data are not the people who make the decisions So we have different types of tailored trainings on on geodata use for for decision makers as well And another one of our flagship products is the population data and heather is our expert From world pop on that. So heather i'll hand the floor to you for these next slides Okay, so i'll be talking today about the high resolution population data That is one of the core geospatial data products that breakthrough produces so grid three Has focused on high resolution data on population in in formats that can be used to gain granular Insights into where population are located both in settings where recent sensors have been conducted and where it's been many years since the last census Now the next slide please And so the population data sets that are produced by grid three are gridded data sets, which means that instead of populations being associated with administrative units Instead population estimates are provided for grid cells within A regular grid and we can see an example here where that has been sort of extruded into a three a 3d Visualization so that the highest spikes are indicating higher population counts Next slide please And so in in this gridded data in gis speak means that this data are in raster format And typically these grid cells are approximately 100 meters by 100 meters So we've got an example shown at the bottom of the screen here. And so you can see the individual grid cells and the estimated population it is shown For each of those with the darker red colors indicated that for a higher population count And the work to produce the gridded data sets is led by world profit university in Southampton And it's grown out of earlier work to estimate population in northern Nigeria And in Afghanistan and given the last national census we conducted there in 1979 So we just go back to the last slide please And so we use two main methods And for producing gridded population data sets and we refer to these using the world property terminology of top down and bottom up So the top down approach involves taking counts of population for administrative units And for example from a recent census and spatially disaggregating these to provide an estimate of population per grid cell And we do this using a weighting layer. So that's derived from a stack of geospatial covariance The bottom up modeling approach instead relies on counts of people in geographically well-defined sample locations So from from a sort of survey and then we develop a custom statistical model to estimate population at the grid cell And including for all unsampled locations. And again, this uses a stack of gridded geospatial covariance to do this And with the development of new data sets over the past few years We're increasingly improving the the accuracy of the population estimates and the availability of new building footprint data sets that's meant to be able to spatially constrain estimates to just locations within settlements And and you saw as uh, Sylvia outlined in one of the earlier slides Some of the other derived work from those data sets and typically The work to reduce the bottom up population estimates has relied on on custom micro census surveys So to get an numerated population within these defined small survey areas an alternative is to use More routinely collected household survey listing data and if these are available and so for exact For example in Zambia recently, we've worked with the Zambia statistics agency to develop bottom up estimates based on household survey and pilot census cartography data in lieu of a micro census While in the Kina Faso We've worked with the national statistics and demography institute and developed a bottom up model model to produce estimates for locations Where during the recent census it wasn't possible to enumerate population and essentially filling in those gaps And the gridded nature of these population estimates means that they can be integrated easily with other geospatial data sets And I'll talk about this further on the next slide And so here we have three examples of spatial units that may be relevant in scenarios related to resource allocation And so for example the vaccine delivery and micro planning And for each of these we can get an estimate of population by aggregating the grid cell values within the area of interest So first an example of aggregating to subnational units So this could be districts or other administrative units or health units And so health zones if those exist And our second example involves calculating population estimates for settlement extent So those that Sylvia mentioned earlier And these are the outlines of settlements derived from building footprints that grid 3 produces And so by combining those data sets we can then have an estimate of population for a settlement And third example here is the estimates can also be calculated for custom units So if you can draw a boundary or specify a distance from a location Then we can also calculate an estimate from the gridded population dataset So these three Examples highlight the real flexibility of the gridded estimates and the potential to be using this at a range of different geographical scales I'm working with these data generally requires some knowledge of gis And grid 3 has had a big focus on geospatial capacity strengthening, but this realistically it's not going to reach all those who could benefit from Using the gridded population estimates in their work. And so to try and reduce this as a barrier to data use We've also made this data available via a web interface Which is on the next slide So this is a WAPA vision So this is an interface where data users can explore The map data and calculate aggregated population totals from the bottom up modeled estimates Which can then be downloaded in a range of formats, whether that's associated with the boundaries or in as a spreadsheet And a user can first specify a country dataset From the left of the screen And then is able to pan and zoom to explore the map so you can see where the higher on those values are And but it also allows a user to calculate aggregated population estimates both for the total population and broken down by Aging sex classes as well. And so this can be for a particular point location Or it can be for an area by by drawing a polygon for a custom area Or a user can also upload a file of boundaries And similarly calculate and population estimates for those locations And just focusing on a technical point from the moment that the bottom up estimates rely on a Beijing modeling framework So this enables Not just a single estimate of population to be calculated, but a full statistical distribution And so this means that a user can calculate rate for single value Which we can think of really is like the most likely value And a range of values representing That the confidence in the estimates and this can be useful for a resource allocation Perspective where if you're interested in reaching 100% of the population And you may want to consider the values corresponding to the upper credible interval rather than the most likely value So we have an example shown For the area outlined in blue on the map With the mean and the range corresponding to the upper lower credible intervals You can see displayed on the right of this room So moving now on To how some of this data has been integrated for other users So next slide please Sylvia And so we have an example here of How some of the high resolution The high resolution nature of the datasets of the population data means that they can be used to gain some new insights into Contextual factors that can be relevant in planning health interventions So for example, in the context of planning codes of 19 public health measures We've used rid of population datasets Along with building footprints to identify locations where it's expected that social distancing would be difficult shown here for the disaster in Zambia And this is calculated from measures of population density And the space available around buildings So the dark blue areas here Have both the highest population densities and also very high building densities indicating It's likely to be more difficult for populations to socially distance in those locations So in terms of planning interventions This could be used along with other data in identifying potential priority populations for immunization And as this potential for faster community spread of diseases in locations And where socially social distancing is all but impossible And so for example, that could be combined with data on the on the age structure data on the population as well And now on the next slide I'm Focusing now briefly on how some of this data can be integrated with VHS too. And so first of all that there are And gridded population datasets, which are produced by world pop available in the DHS to map DHS to maps up and these are available for Countries globally and with an annual time step between 2020 And I'm going to now hand over to Sylvia briefly and he's just going to talk about some new and upcoming work on this topic Thanks Heather Um, so as you can as you've heard from Heather, we do have this population data Um at a very detailed level So at 100 meters by 100 meters So that means that you can actually kind of summarize that data to any boundary Including health facility boundaries So that means that we do now have the chance to get sub district level data Into the DHS too And we have a project that has just kicked off between grid three unicef And oslo in zambia mozambique and kenya where we're looking at exactly that So we're looking at what population data is currently in DHS too How is it being used and how can we add sub district level data? Into the DHS too And how does that for example change our indicators? On the scorecard and so on and so forth One thing that we have been doing is we know that The official numbers that we often see at district level. They don't match the facility Numbers the facility headcount So with this project we are taking The official numbers at the district level and disaggregating them to a 100 meter by 100 meter Grid cell and then re-aggregating them so that the health facility catchment area numbers add up to the official district totals Um, you can still also add other population data in there if you did want to keep the headcounts They're also important But if you needed to match, you know, the official district totals Then that's something that that we can support and we're still looking for other countries apart from zambia mozambique and kenya where we can look at that This is still in its infancy. So we're also looking at the bottle neck analysis app And and trying to kind of find a good measure for the geographic access which we can now do Because we have that Geo data of population settlements and health facilities Yeah, so any questions on that? Please pop them in the community of practice And i'll hand it over to ying tau now to talk about our risk layers Oh, thank you. Heather. Thank you. Sylvia. Um, so i'm going to give a very quick overview of um, our methodology and the risk mapping details so Frame is a geospatial company. Um, that produces hyper local data about people and communities with machine learning Our core output is our raster map data at one kilometer square of a solution That can be easily aggregated to different administrative levels such as district and village And can also be used and combined with other data set, which is pretty similar to the work pop data that i'll just discuss So we can go to the next slide Yeah, so here is a diagram that shows our typical workflow So we generally start by acquiring geotech household surveys and satellite imageries as our modeling inputs So recently we also engage in primary data collection to produce more frequent and insightful information After we get the data, um, we then process them and use them as the input for the machine learning algorithm Basically the algorithm studies the relationship between the spatial inputs and the survey data and estimates values in places Where the survey data is not collected? What we have is the hyper local raster map, which can be used immediately Or be further analyzed to provide insights Um, starting from 2020 we create layers to show different aspects of covid risk and vulnerability in countries of interest We can go to the next slide Um, so we deliver the five risk profiles based on literature Revealed and also consultation with experts and partners This profiles risk elements into five aspects. Um, including exposure risk to covid access to health facility Access to communication comorbidities and socioeconomic vulnerability What we see here is the list of default elements But we often customize them based on the contextual inputs from country teams and local partners And also based on data availability of a given country With a very similar approach, we also develop models to identify color hotspots and areas for a vaccine prior registration So in the color example, we work very closely with first three and ministry of health in san bia on a mapping The output risk maps were included in a reveal report of the ministry of health To complement assisting methodologies for color hotspot identification and disease prevention Um, and in the vaccine space, we follow the WHO guidelines to develop methods for equitable vaccine distribution It has been a model. It has been an evolving model But basically it focuses on different types of elements that focus on confidence complacency and convenience We can go to the next slide Here are just an example. We created for Nigeria So each of the risk composite We actually have a hyper local map for it with the darker color showing the higher risk area So once we have this individual profiles developed, we also create a total risk layer by aggregating Of these profiles to show the total level of risk, which you can see on the top right It actually highlights area where there could be a high need for Vaccine these maps are flexible to use. So on an interactive platform such as a grid 3d data hub or data frame Which is a frame interactive tool User consuming to different locations and also do ranking based on different types of risk levels We also create layers for individual risk layers such as access to piping water Access to soap to accommodate partners need to conduct research and informed decision making If you are interested in learning more about the maps and our projects, we Have a website and also have like demos and health projects For more information. We can share that information in the community of practice channel as well. Thank you Thank you. That's it from our side. Thanks Okay, thank you so much. Sylvia Heathering towel That was really really incredible. So amazing resource that you have available and it's um And it's probably important to highlight that some of the innovations that they presented specifically some of the detail population data Is now actually available in dh is to starting in dh is 2.36 in the maps app you can build maps that allow for the breakdown of sex and age Population data you can render that as a layer in the maps app And I showed a link of it to in the community practice And speaking of the community practice a lot of great conversations and questions already happening there So please do post and my fellow presenters if you guys could take a look at the community practice and and weigh in Um, that would be very much appreciated Okay, so now I think that we are passing it over to um flowminder. So oil it Take it away Thank you. Let me share first my screen Here we go all right So hello everyone and welcome to this part of the webinar where we showcase flowminder's work to support childhood immunization or as we'd like to call it unlocking the power of data for decision making all right, so a quick word about flowminder we are a non-profits organization and it works with around More 30 around 30 people or 30 multi disciplinary staff We can closely with data providers governments and international intergovernmental agencies and we really pioneered the use of mobile operator data for humanitarian and development use since actually height is earthquake back in 2010 So flowminder is of course one of the core partners of the grid tree program And our initial focus is only on operational work and research on use of mobile data for multiple sectors So in detail what we do is everything that is related to data science analysis mobile data partnerships applications in humanitarian and development sectors and we provide a wide range of capacity and strengthening and support activities This presentation actually links with the work that was presented by my grid tree colleagues earlier And describe one of our larger projects where we use key geospatial data for vaccination in the drc in 2021 The president of drc launched the second phase of the plan mashaku The aim was actually to reach at least 75 percent of children vaccinated To do this they reached too many too many partners including flowminder for sports and With this actually we had the necessary data to increase the country vaccination coverage So the grid tree mapping for health projects that's funded by gavi started in june 2020 and focused mainly on providing key geospatial datasets just as maps of localities health facilities health areas in the target provinces for example high resolution estimates of Provincial populations national estimates of population movements as well as some practical tools such as the geo reference the support documents on key geospatial data as well And everything actually was produced to support micro black And other tools in of course to further integrate gender and gender and equity considerations into vaccination programs All these programs and tools Came the associated training to ensure the sustainable integration within the national health system So it's really mandatory We now zoom in on one of these products the population mobility for vaccination Mobile network operators or mno's collect several types of data to record their custom or network activities And call detailed records CDRs cold data records. They are actually the most commonly analyzed for human for humanitarian and development development purposes So using anonymous and aggregated CDR data governance and decision makers can be provided with timely mobility analysis Cold data records are actually owned and automatically generated by mno's for billing purposes So CDRs are produced each time a mobile phone subscriber or simply a user Makes or receives a call Sends or receives an sms or uses mobile data So each call texts And mobile data station creates a record Which includes an anonymous anonymous subscriber ID A timestamp And the ID of the cell tower around in the event So each call detail record appears as one row On the data set So from a CDR data set We can tell the approximate web apps of subscribers or users based on the tower's location And this of course associated with the time of the event that is included in the data set So a CDR data set therefore contains billions and really billions of data Points data points from millions of users cover large geographic areas and time periods So for minors aim is to ensure the privacy of course and We sell abiding to the GDPR regulations and the quality of our final data As a result all the aggregated data is shared and all our methods and are transparent and peer reviewed So we do not provide Data on individuals. That's important This work we make it of course possible to estimate the mobility of The population on a monthly basis and in real time when necessary You can see an illustration of this on the map where you can clearly See the urban centers and the routine back and forth back and forth movement between the urban centers and their catchment areas The advantage of vaccination is being able to observe population changes over short periods of time So due in particular Due in particular to migratory movements But also to see abnormal displacements For example following natural Disasters for example, if there is a flood or something like that And also to identify mobile populations for instance that are mobile Without a fixed reasons who are particularly at risk of being missed by immunization services So the types of the final products are number one monthly updates of population density changes for each health zone specifying for example the The proportion of migrants the proportion of displaced or returned people And the proportion of highly mobile people as well So the origin of each of these groups Provides actually local and national estimates of population change in relation to planning and the annual estimates Also, and it assess actually assess the possibility of using these density changes to update population data in an exploratory research Number two is to actually estimate large population movements and monitor mobile and displaced populations So in order to set up an alert system in case of exceptional events Or to receive maybe the new data automatically without having to wait for the next monthly updates And of course to locate as well and quantify the continuously move the moving population Often missed by the forgetfulness and in this case the vaccination services So the good. So the goal here Is simply to adjust needs and improve vaccinations and vaccination temperatures And the number three is actually the awareness limitations And even network coverage on multiple operators and networks Instability prevents a complete picture of movements Mobile phone users are not representative of the whole population So it is always subject to obtaining the necessary agreements we can carry out Of course telephone surveys, which are then anonymously linked with the location data In order again to estimate demographic characteristics of mobile populations Technically the results can be shown in the form of tables or graphs Thanks to the development of the flow minus software. I think the flow kit To produce these indicators of mobility and change In population density in an automated way from mobile phone data data These results can be of course transmitted automatically through a platform used by health professionals For example the HIC2 After the end of the project population movements estimates will continue to be updated automatically every month If the system is in use and can be maintained And all of this data of course is All of this data processing is done anonymously. I I said it And I'm saying it again. It's done on enough anonymously So at no time can we form either order the test receptions access Individual data to track to track A particular individual that's impossible. This is just big data So in partnership with the university of Oslo and the Congolese's division The national system Flowminder has Has been actually working on the integration of this mobility estimates in the h dh Is so this technical Work in is ongoing still ongoing and the integration is not yet complete But the teams are really positive that the h is the h is to Will be a great platform for the dissemination of these projects products So the platform can also host the associations the associated document documentation meaning that users do not need A training to be able to use the data. So it's really a really good thing and it's really easy to use And we can really apply the use of data here in multiple areas if I say if our site may be You're looking for a placement optimization We're working on urban transport planning or in this case Public health and epidemiology, etc. I said that there are really lots of areas where we can apply this So let's talk right now about the second innovation of flowminder on this project Which is the use of an algorithm They were developed again by us to optimize the number And location and prioritization of proximity services for vaccination purposes So when we talk about optimization It's always a classic spatial problem The questions that we normally ask or receive in this subject are for example, where are The optimal locations in space to situate facilities to situate services to to situate staff how many service points And again staff to add where should we add them? How many existing services? Service points to select or upgrade or downgrade etc. So it's really I'm saying again a classic spatial problem And the answer or the answers to these questions we need to take it to take into account the situation of resources Where they are limited where people are marginalized, etc. So with the limited number of fixed vaccination sites Advanced vaccination mobile on the outreach is a key strategy to reach the children for immunization It is actually essential to locate the vaccination sites where they will be able to reach the most people to make the work cost effective as well So moreover in a budgetary constrained environment, it's important to know the optimal order of of payments of these sites And here i'm gonna again show an example On our request from our Retreat partner season and in collaboration with our casus and the epi We used a our placement our placement and ranking algorithm to suggest the locations And numbers and priorities of advanced sites for two pilot provinces If I could say their name, it's holo mammy and I think tanga Tangarica Of course, there are seven further provinces as part of the mapping for health projects kinshasa, wilu, kasai, oriental, kasai, holo mammy, roko tanga, and holo mammy as well So these results Were included in the geo reference documents supporting microplanned And they were all produced by season and our casus for use by the expanded program Epi some examples of the geo reference documents include a cartographic vision Of course to see the boundaries of the health areas and the health facilities the population estimates the vaccination sites Etc etc And there is also a summary page and a detailed program per vaccination sites That includes population estimates again and other numbers like the frequency of each visit for example So for this algorithm We have certain rules to put in the system For example the server service radius of the health facility that could actually that could actually represent one hour or less travel time A set of population estimates as the 100 meter resolution Again coordinates of the existing health care facilities providing vaccination And if it is Concerned within settlements and that's really important So why should it be constrained? It's simply to ensure that the new facilities were placed within settlements and not in the space between settlements And why is For a prioritizing larger settlements so minimize distance for larger settings as well So even if this implies implies a larger number of sites Is needed to obtain the same coverage as a dickens and unconstrained case, but well, that's what we do We need to So the framework here allows decision makers to evaluate different scenarios To help them determine coverage targets and plan service implementation and expansion Uh, it shows also how big gains can be made with a few service points And the long tail needed to reach 100 coverage stakeholders can see how many service points will be needed to achieve x percent For example coverage or to cover a number of additional people In this instance, I can't see if we go back to our example In whole of my me We and by we I mean the floorminder of season and the actresses We follow the world health organization guideline of an increase in coverage of or at least 150 people This is simply to justify the addition of a new site So this led to selecting 100 and 1,455 locations for outreach for outreach and mobile sites Providing 97.5 percent coverage and In a three kilometers radius, which was deemed satisfactory to the project So the epi tree floorminder and actresses conducted an evaluation mission of the results in the pilot province in october 2021 before applying to the new seven projects What's good here Is that around 90 percent of head nurses in it Demonstrated a good understanding of the documents and more than 85 percent of it is confirmed that they are They have actually adapted their strategies according to the recommendations And what again was other was good is that all of them have confirmed That they want a similar document to help them build their micro plans next year in in 2022. So it's really a good thing And here you can see some of the other examples like on the left where mobile mobile money agency, Tanzania in Bangladesh On the right, we have the schools in Nigeria where we have optimized accessibility, for example Another example here, and I will probably just say the following quotes Free education project will support expansion of radio coverage based on the information provided We really appreciate your support to Sierra Leone But this was very good So as I mentioned before we support the immunization program with capacity strengthening as well And this is offered again by the project as well. And to this date We actually trained more than 500 staff at central provincial and local health agency levels in selected projects provinces You can visit the grid tree e-learning platform for more information if you want You have the websites in front of you So gis and dhi s2 Now I'm gonna say what's the link? Well dhi s2 users can access visualize and download data via the dhi s2 maps app But thanks to the dhi s2s You'll be able to actually combine these maps and geospatial data with official statistics And this can really produce usable actionable useful actionable insights really So again, this really increases the analysis capabilities for example The use of grid of grid population estimates with the health facilities to calculate population Living within certain distances, etc. Now I think we can have a Quick Q&A if there are some questions I think that there has been a few questions. They're posted on the community practice So I think for the sake of time we'll we'll invite everyone to get continue to Contribute to the community practice conversation and I think that we'll have to kind of keep moving along But but oil it thank you so much for a really fascinating presentation It's incredible to see the kinds of tools and Approaches that you all have employed to develop those tools and exciting as you mentioned We we are collaborating To be able to pull in the population mobility data into dhi s2 specifically now focusing on drc But hopefully making this a generic functionality that we can apply in in any country And it's really exciting to see that work moving forward And really expanding the potential and possibilities Before dhi s2 um Okay, so the next presentation is from maha from rafael. He Recorded his presentation in advance. So I just need to share my screen and Let me just make sure that I'm Sharing with the proper Share sound. Let me just get over to the youtube Can I just ask if anyone can hear that? It's okay Yes, it is good. You can go ahead. Okay. There we In the framework of the covax facility. So to reach everyone Wherever they live, it is worth giving a bird's eye view to decision makers namely ministries of health api teams and funding organization on services vaccines equipments and research Taking benefit of the power of location to make sure everyone has equal access to covet vaccines All data are going to play an important role There are web mapping based solutions which display data in near real time at health service delivery points The geospatial platform is aiming at strengthening covet immunization services With the introduction of the web mapping solution The geospatial platform at three embedded master capabilities in the following domains Key capability of the platform Any local planner or health workers is keen to visualize The location of health services in order to gain a clear understanding of all health related operations in the field Thanks to the platform maps can be displayed in electronic image format or printed on paper Maps offer unique capabilities of gaining greater insights using contextual information such as Location of health facilities and settlements in the districts As well as geographical features with roads and tracks even natural barriers That map displays health facilities and the boundaries of the sub districts The map is showing locations according to health facility types such as hospital, maternity, health centers and community health and promotion services as well as roads and rivers You can drill down up to sub district level presently Bechiaco in the district of Ahofo Hano North in the region of Ashanti All the previous expected features are made available at the scale of the sub district As seen before mapping the breadth of potential facilities in the vaccination process Is the first step to answering adequate population coverage But it is also possible to take into consideration the distance and travel time Ideally outreach or mobile team sessions should be planned for all communities within five kilometers of an immunization site Thanks to the information mapped mobile vaccination session could be better planned and managed That map is showing locations according to health facilities type and the five kilometers catchment areas In proceeding with such maps you can easily identify the settlements which are not properly covered In that project in Kyrgyzstan maps were made available with traveling estimates for a 45 minutes transit time from an ambulance parking place Beyond such visualizations health department will need population counts for total population and for each priority group The geospatial platform can easily calculate how many of the population are within five kilometers or a 30 minutes walk of a vaccination site It will be critical to match facility capacity, vaccine supply and population groups across locations to ensure all needs are met You can also display the locations of vaccination sessions The pie charts are displaying the splitting among the three types of vaccination sessions namely fixed outreach and at school the size of the bubble is mirroring the number of sessions Obviously each district exemplifies the core underlying strategy so to reach the highest coverage At sub-district level you can display additional information in regards to coverage rate for particular antigen, vaccine wastage and the number of sessions As a reminder, MAHA developed a geographical information system platform based on online map interface reachable with the web browser with geolocated data-driven functionalities aiming at delineating sub-district areas In the context of Ghana it is currently possible to go deeper into geographical analysis by shifting from 260 districts to 1364 sub-districts Every day you are working towards the goal of immunizing everyone in your targeted population but how can you know if you have reached the targets? Monitoring data is an ongoing process of collecting data, explaining it and really making use of it Otherwise you are not knowing if you are making progress With the visualization analytic tools included into the COVAX Geospatial Platform you can focus on every event that matters most to your vaccine program with monitoring tools so you can make the right decision at the right time Geolocated data are aggregated indicators are available to all the stakeholders at the country, region and district levels Indeed, every organization or expert working for an immunization program using the Geospatial Platform can take advantage of MAHAX and GRAVS to visualize trends such as the coverage rates Firstly, as a kind reminder, dynamic maps displaying coverage rates when used as an epidemiological tool provide effective monitoring and evaluation insights to overcome the complexities associated with the spatial variabilities of vaccine preventable disease The EPI monitoring application designed by MAHA is a user-friendly geographical informational system solution aiming at displaying any indicators tailored to your health program such as coverage rates and district and sub-district levels So you can look at immunization-related data through interactive maps at different administration scales or organization units Let us take the example of BCG coverage rate under one year children You can select the period type monthly or quarterly You can select a starting and ending year the same for the month With the time slider you can visualize the results throughout the month You have the zoom in and zoom out function You can go straight to a targeted district of Chiefdown in the present case Bagruva Thanks to the coded maps and the bar line-based graphs, these visual analytics lead to easy undification of low coverage and unvaccinated concentrations The rates are selected for displaying color-coded maps BCG coverage rate is selected again You can select at legend of your map according to your personal settings by equal intervals or equal counts You can select among a list of absolute figures for the front bubbles The number of BCG doses given is selected Please note that the results are displayed through the line in the graph below You can play with the size or the color of the bubbles before printing the map and the graphs See export function A new set of new indicators and innovative data visualization related to the impact of COVID-19 on routine immunization are made available You can make an appealing comparization of the coverage rate between 2019 and 2020 period to be expressed in terms of monthly change decreased based on red color scheme or increase in green color scheme Above the coverage rate catch-up trend, you can benefit of a line-based graph depicting the cumulated number of children vaccinated divided by cumulated number of targeted children from January 2020 onwards Any antigen can be selected Let's take us the example of Savannah region Any region can be selected as well These points are matching the indicator catch-up coverage rate for the year 2019 The year before COVID According to the curve, the turning point is December 2019 You can deselect districts in order to focus on your targeted choice, namely East Conja The final curve makes sense in regards to the regular decrease of coverage rate between 2019 and 2020 That visualization leads to the obvious conclusion that a catch-up immunization campaign is much needed in that particular district Note that the total number of children to be vaccinated for the catch-up campaign is also calculated As seen before, that visualization is depicting the coverage difference between 2019 and 2020 for the region of Ashanti, Ingana Key Capability of the Platform Turn data into planning and action Share your observations and your analysis Make the data accessible for innovative visualization, clear understanding and improved decision making Thanks to the XGEO special platform, there are other ways to prioritize your activities and control how you want to share the information For example, the experts involved in the supply chain get access to incredibly detailed information about the availability of vaccines, medicines, equipment as well as the staff and services that each facility provides First and foremost, in front of such geospatial platform, both coordination planners and field workers use the same geolocated data so to reduce misunderstanding and boost efficiency Bringing teams in front of such geospatial platform In the framework of the COVAX program, international and in-country stake orders can see telleride information about each country In the case of the demo land, please note that the data are for simulation purposes only and are not mirroring the in-country situation The map overlays are paying respect to the vaccine introduction readiness assessment tool as developed by WHO, UNICEF and GAVI As such, the platform could be deemed as an extension tool You can follow up the following domains of intervention coordination, prioritization and targeting, service delivery and planning monitoring and evaluation, vaccine, cold chain and logistics Each domain has its corresponding features and indicators which are visually displayed nationwide up to service delivery points Let us take the example of the frontline workers in health You can display the total targeted number for each region and you can drill down at granular level For example, Ashanti region You can drill down up to district level so to appraise the initial number of people to be vaccinated by facility type There is a feature to map the country according to the type of health facility For example, you can display by cluster the number of health facility In addition, you have the possibility to highlight all the selected health facilities which are delivering COVID vaccine As usual, you can drill down up to service delivery point with additional information in the right bar such as the catchment population or the number of targeted people Even though COVID vaccine is not yet available at full scale an early implementation of the COVAX geospatial platform suggests that the stakeholders will be better prepared for managing vaccine distribution Okay, so that is the presentation from MAHA And just to quickly point out that as they presented in their video that their application is available in the DHIS2 app hub So that's apps.dhis2.net So you can bring those powerful analytics that they had presented straight into your DHIS2 instance We need to get out of the way for our Francophone colleagues to do this same webinar but in French There's a couple things though that I did want to point you to before we go The first one is that we have now published the DHIS2 roadmap for the next year and a half, nearly two years on our website And you'll see that I'll post this into the chat so that everyone has this link You'll see that a large proportion of our focus of the development of DHIS2 is going to be on local data use that's facility level, district level and micro-planning and vaccine delivery And so that covers not only the analytics tools which you saw a lot of today but it also covers our various forms of data entry our different types of calculations and indicators that we are going to be continuously improving over the next three releases again the next 18 months or so to optimize DHIS2 for micro-planning and local data use That being said, we are of course still working with all the partners that are presented today and I want to thank them again for their fantastic presentations And we are bringing in more of their innovations into the core trying to make them more accessible to everyone And so really a lot of what they presented is kind of the tip of the iceberg You know, we will continue to collaborate and work with them at the university as well as our partners at WHO and GAVI to make their innovations more publicly available and using the large-scale adoption of DHIS2 in dozens of countries as a means to make it available So I think with that we need to go ahead and end the English portion of this webinar And so I want to thank you all for sticking with us Please if you have any more questions, thoughts, concerns, ideas Anything that you want to say regarding the topic of the webinar do share that in the community practice The community practice is here for you to have conversations with fellow DHIS2 users So please do have a look at the community practice link and post there And also for my fellow presenters if you could keep an eye on the community practice and answer any questions that may be coming up specific to yours to your presentation So I think with that, again, thank you very much everyone for sticking with us And I am sure we'll see you all in the community practice And I hope everyone here has a fantastic holiday season Okay, so bye for now