 Thank you for joining us today for the launch of the foreclosure and eviction analysis tool in partnership with data. To start us off, I would like you to imagine something. Imagine that it's April 2020, the COVID pandemic has swept through the country, turned our world upside down. We are hearing that millions of people have lost their jobs. But imagine that our government doesn't actually track employment. And so we have no idea what the unemployment rate is. We don't know who has lost their job, which industries have been the hardest hit and which states are the most battered. Imagine what would have happened to our national policy response had we not had any of that data. And yet this is the situation we face when it comes to housing loss. Homes are the most valuable asset for most Americans. In addition to providing shelter and protection and well-being, we spend the largest share of our wallet on rent and mortgage. And yet the federal government and most local governments collect and share no data on evictions or foreclosures. We have no idea how many Americans are losing their homes each year, where geographic housing loss is concentrated, who is losing their homes and when. I'd like for you to just stop and think about that for a second. Just how crazy that is. How can we make good housing policy, good economic and social policy for that matter without basic data on whether families are able to hold on to their homes. Two years ago, New America and data kind began working on the issue of improving America's eviction and foreclosure data. And to advocate for developing a national and local eviction and foreclosure data infrastructure. Since then, a set of eviction data recommendations we developed together with eviction lab data kind national cities and many other partners have been taken up by HUD in report to Congress on the feasibility of establishing a national eviction database. And that would be the first to tell you that any effort to collect and analyze housing loss data has to start locally. That's because evictions and foreclosures move through our county courts. That's where the data data trail begins. And it's our local leaders, city housing department staff, legal aids, county level housing authorities, local journalists advocates who most need this data and can best use it. For the last 10 months, we and our partners at data kind, along with the National League of Cities, and the Stanford legal design lab have partnered with 14 cities and counties to develop and test feet, an open source tool that allows local leaders to collect and analyze their own eviction and foreclosure data. We're thrilled to release this tool to the world, and hope that you'll use it. I think I neglected to introduce myself at the outset. My name is Julia Panfill. I'm the director of the future of land and housing program at New America. I'll now hand things over to my colleague, so be has I know by who led the development of feet to tell you more about how we got here. Our partners at data kind, Mallory chef and my new Sharma will demo the tool, including its outputs. And after that, we'll move into a panel discussion, featuring leaders from three of our city partners, Hayward California, Hampton, Virginia and Tucson, Arizona, as well as the National League of Cities to discuss how cities are using this tool in real time to improve their understanding of local housing loss. And with that, I'll hand things over to see. My name is to be his anal by and I'm a senior policy analyst at the future of land and housing team at New America. As you will have mentioned, we wanted to provide a little bit of context for what precipitated the development of the feet, the foreclosure and eviction analysis tool in the first place. Eviction and foreclosure is the most common form of displacement in the US in a typical year and estimated 5 million American through eviction and foreclosure, and that's just what's being documented. Displacement through informal means are not recorded and are thought to far exceed instances of formal housing loss. So for this reason, and over the past two years new America and data kind have partnered to better understand housing loss across the country. We first began by collecting as much eviction and foreclosure data as we could from county courts across the US leveraging this data to generate insights about who's being displaced and where. What soon became apparent, however, is that without a coordinated or standardized data collection mechanism of eviction and foreclosure data, developing a comprehensive understanding of what is happening is challenging, if not impossible. So in addition to analyzing evictions and foreclosures and select localities where housing loss data is accessible, including seven counties across the Sun Belt. We also have been working in partnership with data kind and a lot of other organizations like eviction lab and national low income housing coalition to develop a set of recommendations that build out local eviction data infrastructures that can feed into national databases. So through this work we've really learned a lot about the housing loss data landscape. Next slide please. And as I mentioned one of our key takeaways from this work is that the access access to housing loss data is one of the biggest barriers to understanding more about displacement for cities and counties. Next slide please. So our previous research finds that one third of US counties have no access to eviction data. And that's to say nothing of whether the two thirds that do have access to data have access to data that is high quality and comprehensive. We know from from our prior work that cities and counties face either incomplete or partial data and data that varies widely and whether it's standardized digitized or whether it's publicly available. But still a majority of counties in the US do have some level of access to eviction and foreclosure data through their county courts. Next slide please. And we know how crucial having eviction and foreclosure data is in preventing housing loss, especially in ways that are responsive to local demographics and local housing loss. So without this data we don't, you know it's difficult to know where housing loss is most acute who is being evicted or foreclosed upon whether evictions or foreclosures are rising or falling over time. In addition to things like how much back rent is owed. And if during the pandemic for example if evictions and foreclosures are continuing despite the moratoriums. And all of these things make it really difficult to know whether. It makes it hard to know whether how to target rental assistance and how to conduct outreach for for things like eviction diversion and other upstream pre filing programs. It also makes it difficult to track and understand how evictions and foreclosures are perpetuating racial inequities and to see whether measures put in place to curb this are actually effective. And ultimately it just, it makes it more challenging to just keep people housed. Next slide please. So the catalyst for developing the data tool was really in response to this pressing data need at the local level, in which those on the front lines of tackling these housing crises lack not only the data but often also the capacity, including the resources to be able to incorporate into decision making in a, in an ongoing basis and not only for, you know, a point in time one off analysis. So building off our past work, we partner with data kind and the eviction prevention learning lab which is co facilitated by National League of Cities in the stand for legal design lab to develop this foreclosure and eviction analysis tool which you'll be hearing a lot more about in just a second but at a high level it's an open source data tool that allows local leaders to upload their own eviction and foreclosure data and generate a series of analyses that is designed to answer a set of questions. And it can analyze up to three types of housing loss, data on evictions, mortgage foreclosure and tax lien foreclosures. Next slide please. So a critical component of the tool development were was our partnerships with 14 cities and counties who are proactively working to improve housing loss data and analysis in their home communities. These cities and counties are all, they all varied in their access to data and in their analytic activities to date, which provided us data kind in New America and developing the tool with a good range of context to understand how the data tool could work for and could work for and be generalizable across as many jurisdictions across the country as possible. So we'll be hearing as Julia mentioned from three of those cities as part of a panel discussion. But there were many more cities involved in this as well. And you can see on our website a list of all of those. But before we get to that panel discussion I will pass it off to my colleague Mallory to provide an overview of feet. Thank you so much to be and good afternoon everyone. My name is Mallory chef and I'm data kinds portfolio manager. I'm really delighted today to be sharing more information about the feet tool for a little bit of context, data kind is a global nonprofit organization and we really harness the power of data science and AI in the service of humanity. We attempt to bridge the gap between technologists and the social sector. And we bring together talented pro bono data science experts with social impact organizations. And so if we go to the next slide, we'll see here that for the creation of the feet tool. We actually worked with a small group of incredible talented pro bono data scientists to really drive this work forward. And thanks to whom this idea became a reality. So look at as to be shared earlier can really work with three types of housing loss data evictions, mortgage foreclosures and tax lien foreclosures. As you can see here, the functional design really highlights the complexity of the various modules in ingesting, transforming, and then providing key findings on housing loss for these local counties and communities. And feet, I'll walk you through these four primary modules of functionality. The first module is the load module. And this helps accurately ingest the data on evictions and foreclosures that local communities can upload directly into the tool. The module is important at it really ensures that the subsequent modules run smoothly and can really generate the impactful analyses and visualizations. The second module is the transform module and this gives a transformation from the given address data from the evictions or foreclosures and translate this address data into street addresses into. I'm sorry it translate the streets addresses into the corresponding national census tract ideas ideas. And this really allows that the unit of analysis for visualizations and outputs to be at the census tract level. The third module is our analyze module, and this module aggregates and analyzes the eviction and foreclosure data against a set of 60 American Community Service survey variables. These variables were chosen from our previous subject matter expertise and collaboration with New America, as these variables were most likely to uncover important trends regarding the sub populations that were most and least vulnerable to housing loss. The output results of this analyze module, which will review in the next few slides, and which will also present to you directly through the live demonstration include a time series chart correlation results and an output file to facilitate any further analysis that the user would like to do. As our fourth module, and if we could just go to the previous slide apologies. The fourth module in the functional design of feet is this visualize module, it really produces a package file that can then be imported into an open source GIS software, such as QGIS. And this really provides custom housing loss and demographic maps that at a census tract level provide visualizations to the local cities and counties that are using the tool. And so each of these four modules really perform critical steps in the data processing and analysis pipeline. Next slide please. Just to share at a high level some examples of the tools output which are live demonstration will also further illustrate. We use here at 2017 data from Hillsborough County in Florida. As we can see here, one of the tools outputs are time series plots. And these really provide information to localities so that they can understand when over the course of a year evictions and foreclosures are most likely to occur. Next slide please. With this second tool output which are cord correlation analysis, we're able to provide cities and counties an opportunity to better understand who is most at risk of eviction. As you can see here we're using these 60 American Community Survey variables to really demonstrate the richness of understanding of risk around eviction and foreclosure. Next slide please. In this last slide we can understand at a visual level, the maps at a census tract level around housing loss to really understand where in a city or county evictions or foreclosures are most acute. And so these are just examples of the tool output, but I'm actually now delighted to hand it over to Manu Sharma, who was a team lead for our incredibly talented pro bono data scientists. And he'll walk you through the use of the tool and demonstrate the outputs that are generated from each module. Manu, over to you. Hi, thank you Mallory. I really appreciate it. And it's an honor and a privilege to be here introducing and demoing this really exciting and we believe really important tool which again I'm really glad I and the team were part of. So let me go over to a couple of things here. First things first, the best resource to understand what this tool is what it does how it works what do you want to do with it and how you want to run it is this very beautiful and very comprehensive and very detailed guide put on the new America and data kind team on how to use it right so like more than anything else I'll be able to show you like this is your first point of reference on how to use the tools. But the point of doing a live demo here is to is the hope that like just demoing it live makes it a little bit more vivid and meaningful for everyone who was on the call. I'll go through a demo but again I mean anytime you have questions you can you can go through this guide against really nicely written those four steps tells you how to collect data how to assess the data format data for the tool, how to run the tool and to read the outputs, all that right so again I'll go through them very briefly but basically again all the details that you need are here you can click into these things and again see anything you need. So the first things first and it's a software tool, so it's built on a set of code, and where do we find the code. The code is in a GitHub repository. And if you just, I mean again the user guide lays it out tells you exactly where to go but in case. If you want to search from scratch you go to get up.com and you can just put in something like New America housing loss and it comes right up. So again just to really quickly repeat this, the user guide is awesome. This is your first point of reference has a bunch of these sections with all the details to go to this first, but I'll show you how to look at the tool. So first of all, let me go back to get up GitHub is where the code is stored. I am going to put in a new America housing loss. Here and you can search and find this repository which is new America housing loss public right. That's where all the code lives if many of your data nerds or coordinates you can go in there and actually look at all the code it's open source. And there's a bunch of instructions here. Again, all of these are again listed in the user guide but the point I want to make here is that these are here for a reason that it's like the way the code works it's it's pretty important to actually follow these somewhat rigorously so that you can run it correctly and you don't run into unexpected errors so again these are the instructions from the tool how to install it how to download it how to format your data and where to go to again download libraries and such different operating systems all that so it has everything here. Next, let me show you a quick data set I mean just to again make it a little bit more meaningful as to what your input data looks like. So, basically it's something like this. This is a set of it's a set of fake data by the way so I'm allowed to show you actual addresses real addresses normally that would not be good. But the main point again this fake data I can show you the addresses but basically the main point is that anything you you input into the tool the basic criteria are an address and a date. Those are the key things we look for so that we can plot the time series, we can geocode data and so on right and this is a set of fake evictions data. So basically again this at a raw level each of the evictions in the in the in a certain like you know geography it could be a city county and whatever is relevant to your use case. And in terms of how it looks in your file structure typically what we recommend is that you have a specific directory. And the structure of the directory again is quite important. You should only have the input data in the directory and you know you can have evictions data mortgage foreclosure data and tax lien data as again Subya and Mallory previously mentioned. So these are three types of data you can have it's okay if you don't have all of them, but you need to have at least one of them for the tool to work. And that's that's how typically set up the input data directory. And when you run the tool, all you do is point to this directory as like telling the tool saying hey here's where you can find my data. Now run the code and produce my output data which I'll show you in a little bit right. So just a quick now overview of how the tool actually works how it runs. So this is my screen at the very top you'll see a bit of like you know it's a long path but basically the important part is you do want to be in this CLI directory to run the tools. That's pretty important again the instructions also say this both in the user guide and the GitHub people but again, be careful about this. And the command you're going to follow is going to be somewhat I'm on a Mac so I have a slightly different command on Windows you might have a slightly different command again all this is listed in the user guide and the instructions but basically the command says run this file which is the main file for to enter into the code load data. So I'm pointing it to the exact path of my data which is which I just showed you in my sort of Explorer window right and so it like it was in this data kind data slash input data, and again this hyphen at the slash the end is kind of one. So once you give it the input power that starts looking into all the directories recursive through all of the various kinds of data reads all of them, read all the files and tells you what it's doing that I'm again that there's a lot of output here that you're going to see in the screen and it's quite design because we want you to know exactly what the tool is doing. Again, first of all, it's transparent and secondly, it tells you any errors that you have in the data or that you might have encountered and running the tool so that someone can potentially debug it right I mean if it just fails and fail silently would never know what what happened right so. Again, that's why by design we have a lot of output it tells you like shows you a sample record it shows you if there's duplicate data in your records. It tells you if there's a certain columns are missing I mean they can potentially be and it'll try to find some alternatives but again it's recommended that you do follow the guide somewhat rigorously so that you have the best chance of running it successfully filters out like values are too old goes through again the inventory eviction directory goes through foreclosure same thing can power much data you allow it will go through all of it and just run through it so it does some data validation for all of that stuff up here is all about data validation. Once the data validation is done. Now the point is, if you have addresses right you do want to geocode the addresses so that you can convert an address to the census tract level and that's the sort of unit of analysis for for all of the data. So that's the aggregation analysis unit. So, as it says I mean just again, it standardize the addresses so that you have the best chance of hitting the census geocoder and getting some results back because I can add that a malformat to sometimes fail out. So it does some address formatting and if you have a bunch of strange addresses that the standardization part the code can't look at it'll just tell you that these are the addresses that are not like you know standardizable. So just look at them, correct them if you want to rerun it right so again just this data set I used to run this tool is had a bunch of these strange ones but it told you exactly what is again a lot of output but by design. Then it finally geocodes them and geocoding is a relatively time consuming process depending on how many records you have so there's nice progress bar here to tell you how much time is left and how much time is elapsed. So as you can see for the data set I ran it ran in 32 minutes right so it had quite a few records so it and it told you how many records it actually geocoded. And then the same thing, when it had eviction and for closure to look at what did both of them and series. The same thing there and then there's a bunch of different methods to try to geocode data that that's all again module one in some sense right that is all the module one part the melody showed you right now the thing is okay well and then module two is a geocoding part so the module one validation the data standardization geocoding and then getting some of the ACS data to append to it as the second module right that's a transform module. So for ACS data again that's another census API that it hits to get it. It, like, you know, if it this is an error it'll fail out here sometimes the geocoder is down etc but again it gets all the data and then it creates this housing loss summary by appending all of these 60 variables plus all the all the other geocoder data into a single data set that you can look at and kind of customize our calculations. And then finally, module three analysis right I'm that it shows you all the analysis results saves all that all the plots to the output directory I'll show that to you in a second. And then finally module four for visualization it gets this geopackage file which is this GIS data import.gpkg which is basically the file you can import ArcGIS QGIS and like any sort of mapping tool and like visualize like how your housing loss looks right so that's basically an overview of like the run out personal tools from beginning to end again very copious outputs by design hopefully they're helpful and just letting you know that it all worked correctly right and if you get to this geopackage point yes congratulations it did work. And then finally just to show the outputs real quick. Once you have the input data directory I mean again typically this output data directory is created after the whole tool is run and done everything it needed to. So we'll quickly click into this directory and show you everything so basically produces a few things it's full data sets is just a bunch of. Again it's all your data basically geocoded like so essentially all your geocoded data the raw data is going to be here. So if you want to do something custom with that data this the tool does not do like that's like your geocoded data right there right data summary same thing so you have an ACS data dictionary the ACS variables are a little inscrutable by name so. We have a dictionary that we output and then the housing loss summary which has all the all your geocoded at the census track level joined with the ACS data like all the ACS variables again any customers you want to do that's why you want to do it on. Then it has the plots which is what Mallory showed so the time series plot is again this is what it is and this actually is real data by the way. I can't tell you who it's from but it's it is real data. And as you can see it's it's very useful and if you can imagine right this is 2020 March and April's like you know there's. This sharp drop off in eviction for example is going to be at least partly driven by the CDC evictions moratorium right so you can see some of these trends, they are going to be quite useful to see I mean same thing on for career see a bit of a dip here. So, again that's that's basically the intent of this time series plot. So you have the some of these correlations that Mallory showed you there's there's a lot of stuff here but I'll show a couple of important things. So again, Mallory showed this to you, which is your foreclosures. So the housing laws collared with all the ACS types of variables you've given them descriptive names because the ACS names are less than descriptive, let's say. So you'll again see for your geography, what are the biggest correlations positive or negative the negative ones here positive ones here. The second part of that is to say okay it's not just the correlations that matter. Like it's also the context of the right so what this does is this contextualized correlations tells you. I mean you, you were shown some previous work that data kind of, and New America done together on other other geographies. And what this does is puts your results against like benchmarks them against like what we've seen some previous data so again there's nothing wrong if your red dot which is your data is like outside of these like box and whisker plots. That's totally fine I mean that is your data it is it is your it's unique to your geography that, but the point of this is that potentially it gives you something to look at and say oh is do I potentially have a data issue or something very unique to my geography that might be different than what other people are seeing right and then that's that's valuable information as well right so that's that's why we produce this contextualized correlation plot. And then finally and this detailed results like this is like I'm not going to go into that but this is like literally the plot of all correlations are all very all ACS variables against your data. Again which ones are positively negatively correlated. So again every single data point that you had in the data is basically going to be one of these plots right so again I'm not going to go into that, but that's basically how your outputs are going to look like. And that's all from me, in a sense, so I'll stop here handed back over to to be her and. Yeah. And hopefully that was useful. Yeah, thank you so much money for that comprehensive overview. So just really quickly before we head back or head over to the panel discussion I wanted to share how you can locate some of the resources, the tool itself and then some of the resources that on you shared. Next slide. Angela. Thank you. So we will share this PowerPoint on our on our web page along with the live recording and so these are all links here but there's a publicly available GitHub link, where the tool is actually stored and where the tool can actually be downloaded that of course there's the comprehensive step by step user guide. I will just reiterate that as I'm not, I don't have any Python background or coding knowledge and so that user guide that we created is really intended to not only provide documentation for the tool but also for anyone who might find Python or using a tool like Python or coding. It's complete with screenshots for every single step and really, when it comes down to it after, after downloading formatting and storing the data, according to the guide, the guidelines in the tool, there's really only three steps that need to be done in Python which are mostly just navigating to where the tool is stored on your computer, navigating to to that input eviction and or foreclosure data on your computer and then running the tool against the data. So really, all of the code and output that Manu is sharing on his screen is all generated by the tool itself, and it's really meant designed to sort of easily automate those functions to create these outputs. We have a FAQ page which is at a high level sort of just answering questions, you know, what is the tool, who's it intended for what kind of data do you need, sort of as a resource to be able to share with, you know, local partners who might be interested in a tool like this. And then lastly, through partnering with partner sites. We also developed a blog series called data to drive housing loss decision making which is available on our website where we sort of document insights from both the partner sites but also just our other work in this area. And you can find things like a brief we wrote a brief on techniques to measure informal evictions, for example, and we hope to really continue to build out that blog series to share insights on creating this local data infrastructure. So, before moving on to the panel discussion and passing over to Julia. I just wanted to reiterate that one thing that Manu mentioned that's important is that part of the reason to create an open source tool like this is so that data can remain confidential both eviction for closure data. You know where the tool can be downloaded on your on your computer and your local data can be uploaded into that tool so all ensuring that all local data is remaining confidential. And with that said, you know that you saw the outputs that Manu and Mallory both shared and the hope here is that we can build upon this tool to conduct additional analysis to that we know is at the forefront of cities and counties. And that needs and questions as they attempt to prevent evictions for closures. So, yeah, with that I'll pass it over to Julia for our panel discussion. I am so pleased to welcome our panelists. Lauren Lowry, who is the director for housing and community development at the National League of Cities, and the lead on the eviction prevention learning lab. Christina Morales, who is the housing division manager for the city of Hayward in California. Sarah Lanias, who is the community safety health and wellness program director for the city of Tucson, Arizona, and Lauren white, the chief neighborhood development specialist for the city of Hampton, Virginia. Welcome to all of you. Please feel free to turn your videos on. So we will have a about 35 to 40 minutes of discussion, followed by audience q&a. So please, you know for the audience do feel free to submit your questions using the Slido function on your interface and we'll be posting those questions to the panelists at the end of the discussion. I will start with Christina, Sarah and Lauren white. It would be great if you could just briefly introduce yourself and tell us a little bit about your motivation for engaging in the National League of Cities and Stanford legal design labs, eventually eviction prevention, learning lab, and then also the development of the feed tool, and we can go in that order Christina Sarah and Lauren white. Thank you, thank you, Yuleen. I'm Christina Morales housing division manager with the city of Hayward and thank you for having me here today. I was motivated to participate in the eviction prevent prevention learning lab because we have a recently adopted rent stabilization ordinance and trying to help community members understand their rights under the ordinance we found challenging and the eviction prevention learning lab would use that could provide us the support we need to better help us understand how to reach out to those individuals. And specifically for the housing loss data tool when we had updated our, we recently updated our ordinance. There was a constant conversation about meeting more data from our council members from the community from people who wanted housing policy changes to groups that didn't want it to change. And there was just a complete lack of data. And what we're hearing from the community was a plea for help that the housing prices were too high they were facing eviction. And what we're hearing from the landlord groups is that they needed to be able to support the properties have sufficient revenue to cover their costs, and it had a right to a fair return. What was happening is the conversation was turning to one where it was a conversation about having isolated tenant incidents, incidents from the landlord groups and losing that plea for help. So we really needed data that would demonstrate that there was a pervasive problem that needed a legislative response to help with. So next and again my name is Sarah Lonnie us happy to be with you all this morning so you know for the city of Tucson and our partner Pima County participation in the eviction prevention learning lab made a lot of sense for us because we had been observing ordinarily high eviction rates long before COVID. And it was kind of a challenge that community members were facing but there was no really strong coordinated effort around that so through the learning lab and then through the partnership with data in New America, we were able to really think through how can we better utilize court data and make it more accessible and so just to kind of share it's not as though there weren't already folks in our community, specifically partners at Pima County that were looking at that court data right and looking at it spatially to understand where evictions were happening, but because and if for those of you who are tuning in if you didn't catch it, one of the things that the tool does is it makes it much much more efficient to geocode that data so where we were previously having staff members at the county spending an extraordinarily long period of time kind of trying to figure out addresses as they were put into court records right getting those fine tuned so that it could be geocoded and then could be looked at on a map so that we could understand where evictions were happening. It was a really intensive process and so that created some barriers quite frankly to being able to really look at the tools spatially, and to do some of the additional additional analyses that were needed. And so, you know, partnering with data kind of New America on this tool was very very helpful for us in terms of identifying an additional, an additional resource to actually make that much more achievable right and to open up those pathways for our partners working on eviction prevention. So, thanks. Good afternoon, my name is Lauren white and I am the chief neighborhood development specialist for the city of Hampton, Virginia. We are a coastal city in Virginia of about 135 people we're right on the Chesapeake Bay. We were motivated to participate in the eviction prevention learning lab and the partnership with data kind, because in 2016 eviction lab released a report and Hampton, Virginia, along with several other Virginia cities were in the top 10 list of highest evictions across the whole United States. So that sparked a conversation both at the state level and the local level across Virginia about what we could do to reduce our eviction rates in Virginia. And so this partnership with data kind has been very helpful for us to get a handle on the first of all the data that's out there, and how we can utilize this data to lower the eviction rate. Thank you so much. And now, Lauren Lowry, I will turn it over to you. And perhaps we should have actually started with you to introduce the eviction prevention learning lab. But nevertheless, if you just tell us a little bit about the eviction prevention learning lab, what it is and then talking across the ETLL cohort, what have you seen as cities most immediate needs when it comes to eviction prevention? Yeah, thank you. Lauren Lowry, the Director of Housing and Community Development at National Lead Cities. The eviction prevention learning lab is a peer to peer network where we're constantly looking at best practices and policies as it relates to eviction prevention. We drive into things from eviction data, to eviction diversion, to landlord and tenant education, to communication outreach, to court support, where we're kind of looking at evictions comprehensively. And looking across the 30 cities, as it relates to the most immediate needs, as it relates to eviction, that cities are grappling with on a daily basis. I think one is continuing awareness and education, on how to really target outreach to both renters and landlords, and what does that mean using traditional means or non-traditional. Also how to target rental assistance to ensure equitable distribution. Building is sustaining a eviction data infrastructure. I will also say funding because the main question is, after all these emergency rental assistance programs are gone, how can we continue to funding the infrastructure that has been built during the COVID-19 pandemic? Thanks, Lauren. And turning it back to the cities, Christina, Sarah, Lauren, White, what are your, to the extent that you do touch on this in your introduction, what are your city's most immediate needs related to eviction but also foreclosure prevention to the extent that you're working on foreclosure issues? So I think our most immediate need, as Lauren mentioned, is that as the rental assistance subsides, we have to figure out ways to support community members. It's becoming more clear that people are being evicted because they don't have enough income to support their rent payments. The city of Fort Worth, California have been skyrocketing and its housing has become unaffordable to the community. In terms of foreclosure, we are also, we have implemented another program that's going to provide cross mitigation counseling to homeowners so that they can avoid foreclosure. We're anticipating that while in California there is a program available from the state that will provide financial assistance, not everybody will qualify for that if their servicer is not participating in it. So we're concerned that many low income homeowners may lose their housing and with rental prices being as high as they are, it's hard for them to recover and stay community members in Hayward. I'll jump in here and echo exactly what Christina shared and then also just share that, you know, as part of our participation in working with the tool, we were able to draw upon our eviction data fairly easily and so we ran the tool using our eviction data. We still have yet to do that with the foreclosure data so that is definitely a need for us to really kind of get a handle on how, how foreclosures are shifting at this time. The other piece I think that we have been trying to make sure that we're getting a better handle on is pulling in data from where households that have received assistance through ERAP. And then also, you know, while we were working with the tool, we were also launching an emergency eviction legal services program really through our partners at Pima County. So wanting to understand kind of the geographic impact of both of those programs vis-a-vis, you know, trends around eviction over the last six months and then moving forward. Our most immediate need has been using the data to target people who are experiencing eviction, but doing intervention before it reaches the court system. We have really honed in on the part of the tool set that identifies potential actors or identifiers that can be used to help target our resources and outreach to help people before their evictions reach that court system. Thanks. So we'll move back to the city panelists in a second to talk a little bit more about the actual process of getting the data that you needed in order to use the feed tool and some of the outputs that you were able to generate. But before turning to you, I wanted to turn back to Lauren Lowry and ask a more general question, looking across the cities in the EPLL cohort, you know, eviction data is kind of a subset of the issues that the cohort is working on. But from what we've seen, what are the most needed types of eviction data that would be the most useful for the cities in your cohort in order to make data-driven decisions around eviction prevention? Yeah, the types of data I would suggest would be historical. I think collecting historical data allows cities to establish a baseline of what eviction fountains are given in a week, a month, and a year typically. I also think about geographic data to display the addresses where each eviction is occurring to identify hotspots and to be more strategic in deploying resources. But also say demographic data is important, that way programs can be designed equitably and we can really begin to think about how renters as well as landlords are assessing the funds and programs available. But also what is important is program metrics, right? We have eviction filings that have been filed and that have judgments, but it would be really great to tie in program metrics from mediation programs, housing navigation programs, right to council, and emergency rental assistance programs to have a more comprehensive view or outlook as to how evictions is impacting local communities. Thanks, Lauren, for going through some of those points. So I want to turn to the city is to talk through the actual use of the tool and as we found in our own work, you know, before developing this tool, we tried to source eviction and foreclosure data ourselves, and it was such a nightmare. Like it was just, you know, we tried to source it from, you know, maybe a dozen different county courts and it ranged from being, you know, relatively easy in a few cases where that infrastructure existed to being just incredibly time consuming and challenging and figuring out how to get our hands on data that was purportedly being generated, but wasn't being aggregated into a database wasn't being shared, you know, just dozens of phone calls. I remember people physically walking down to courthouses with checks in hand in the middle of COVID. I could go on and on but instead of me going on, I'd like to turn it over to Lauren White and also Christina, you, both, you know, both the city of Hampton and the city of Hayward didn't have eviction data on the ready. You needed to go through the process of obtaining this data from the county court or somewhere else in order to test the feed tool. So can you share a little bit more about that process. What barriers did you face and how were you able to overcome them and what lessons are you able to share for other cities who might be grappling with these same challenges. So in Hampton, we receive our eviction data from the local sheriff's department. And the reason we decided to use that as a source is because that represents the actual evictions, you know, who have been served that written eviction and are in that two week period of moving out. So we use data from our sheriff's department to measure our eviction rate. Prior to this process, our community development department and our sheriff's department were not interchanging that data, it was just sitting on a computer in the sheriff's department. They weren't measuring anything. They were just getting their orders from the court and then going and serving them. So part of using this tool was coming up with a process to share that data between the city and the sheriff's department. So we could effectively use this tool and really set up a process for, you know, sharing the data. And now, as a result of this process we actually share the data quarterly so I get a report from the sheriff's department. And I sometimes have to manually go in and make sure things aren't duplicated so there's still part of the process that we can improve upon. And then I share that with our IT department and they are responsible for mapping and, you know, putting together reports. And so that has been one of the great outputs of this process is that we do have a process now for sharing and analyzing our eviction data. Yeah, so for the city of Hayward, I mean in California, California law limits what is what kind of information about evictions is publicly available so most of the county courts don't have any data available. Initial requests for information from the city were, you know, turned away you have to know the parties of the eviction in order to be able to request the information. So it seemed like an impossible task to be able to get the information, but being part of this process and knowing the opportunity we had. We started the process knowing we had some records based on notices of terminations that were being filed. We wanted to make one last attempt to see if we could get better data and so we reached out to our board of supervisors supervisors office that represents Hayward and asked them if they had any connections in the county courts office. And they connected us with the executive officer, who then explained the administrative process for requesting public records through the courts and what the limitations were and what specific data we could ask for, and who the proper contacts were. So we were able by going above our above to our county board of supervisors we were able to leverage the right contact to be able to access some walls still limited access information that is definitely showing us kind of the volume and magnitude of evictions in Hayward and as Lauren kind of pointed out looking at that historical data, knowing where it was in 2019, we can see that it's it's dropped in during our moratoria, and then trying to make sure that it doesn't go up back up to those high levels once the, the eviction moratoria have ended. Thanks so much Christina and Lauren. Sarah Tucson already had, as you mentioned an existing pretty clear pipeline for accessing eviction and foreclosure data. So in the case of Tucson. What were some of the barriers that you based in analyzing this data and incorporating it into decision making and how are you able to overcome some of those barriers. Yeah, thanks for that question. You know, one of the biggest barriers for us and I feel mildly silly saying this but I think there's a lot of people tuning in who can relate to this which is that we had an all hands on deck situation throughout City of Tucson and Pima County and folks who were really focused on eviction, they were on the front lines, providing eviction prevention services and so being able to really truly just take the time collectively to review the data to ask additional research questions so that we could go back and do further and then critically to be able to implement right to act on those findings was slowed down a bit. I think for good reason, right, because we were able to actually provide emergency rental assistance and provide emergency legal services to really change the experiences that folks were having in eviction court proceedings. But so we are still in the process of really going back and doing more of a deep dive there and bringing power together to really make sure that we aren't having any additional blind spots right that we're that we're really able to get the most learning out of that data and then to think strategically about how to act moving forward so that it's an ongoing process. For sure. Great. Thanks so much Sarah and as you can see from Sarah Lauren and Christina's responses. And this is sort of what we found when we were pulling together the cohort. Cities kind of lie along an eviction and foreclosure data continuum from some cities who have no access to any data to other cities and counties who have quite sophisticated access and analytic capabilities beyond what feet is able to offer and what the feet tool tries to do is to wherever you are situated on that continuum to provide you some tools to analyze the data that you do have and help you obtain maybe the next level down granularity of that data. So, so turning back to the city panelists, one kind of last question in the weeds before we zoom out. All of you were able to run the feet tool on your eviction or foreclosure data and I'm curious. What types of insights were you able to gain about your cities evictions and foreclosures with the help of feet. And I know it's still early days. We just released a tool but how are you planning to leverage those insights in mitigating evictions and foreclosures. In Hampton we were able to use the tool to identify areas in our cities with the highest eviction rates, and that has been really helpful again in targeting outreach and services to make sure that they reach the people who need them the most. It's also been helpful to identify landlords who may be in need of assistance and may need help navigating the court system and doing additional outreach to landlords because sometimes they don't know or understand all of the programs that are out there that can help them before they file this eviction. So it's just really been helpful in identifying the areas and then opening the lines of communication with both the landlords and the tenants who need that information. Thanks I would piggyback on that and you know one of the things that we were able to do by, you know, looking at a broader swath of data right so part of what the tool allowed us to do was to go back further and so we had some baseline to compare to. And so, you know, to echo what Lauren was sharing, it allowed us to be able to do some very strategic actually like apartment complex based resource fairs, geographic focused resource fairs around eviction prevention and a host of other kind of related services. So being able to look at that before and after I think also helped us to fine tune what some of the needs might be and to, and to make sure that we were making those available so that was, I think very very helpful for us. Another thing that I think it has helped us with a bit is just to get a sense of where there's some additional data sets that we could pull in for even greater learning so again, where are we seeing high rates of eviction but we're not seeing upticks of emergency rental assistance right you know is that because there's a landlord that's not accepting the assistance or you know what is that so it's so I think as we're able to pull in some additional data sets, there will be even further learning there and so going through this process has really helped us to identify what some of those additional data points are. Thanks. Okay, we're very similar to the other responses but I think the importance of correlating the eviction data with the demographic data really paints a picture of what the risk factors are for people to be subject to eviction or foreclosure. So helping us to understand what those risk factors are helps helps staff to develop policies for councils or board of supervisors considerations that would address the specific needs of the community members who are at risk, versus trying to do a blanket policy that may or may not help them. So it really helps us identify this is the population that is at risk, and these are the strategies to help them. Thank you. I'll start with Lauren Lowry, maybe speaking, generally across what you've seen from the cohort and then turn it over to the representatives from the cities for this next question. What else do you wish that the could tell you that it doesn't occur if you know so what are the either types of data that you know you'd be able to collect or, you know, kind of linking with other data sets I know several of you mentioned that, you know what other functionalities we're hoping to keep building and developing this tool what other functionalities would you like to see. I think the integration of program metrics would be really great to really see who is accepting them and who is not. And I really liked how Sarah pointed out. We don't know that the program exists right and so again emphasizing that outreach and awareness but also informal evictions would be really great to because what we don't know we can address right so those are some of the things I would like to see in the tool. I would have liked to see other housing or affordable housing metrics that would help us to understand our eviction problem or even drive our eviction problem, we all know how the housing market is right now. So having other metrics that touch on our how our housing market I think would be very helpful. I love that you said that morning sorry, you know I hadn't even thought about that but you're right that that would be very very helpful to look at. One of the things that that myself and and the team that I was working with that we kept wondering about was kind of how does Tucson stack up with other, you know, other cities, you know nationwide so wanting to just be able to look at the tool some kind of baseline data, you know based on other geographies so that we kind of know I mean I hate to say it but so we know kind of how we're ranking there, because that can be really helpful in terms of motivating the local leaders right to act. So, you know that that would be very helpful. You know one of the, one of the interests that we also had was to be able to look at the eviction rate through the tool while accounting for housing tenure type right so accounting for the percentage of households that are renters. And then we get a little bit more of a of a fine tuned understanding of what that rate of eviction really is right based on a census track. And then, you know, another kind of wish was, you know, could we get some additional data at the block level to work with. So, yeah, those are a few things. One is there's so much data and I don't think we've gone through it all but what comes to mind in addition to what my other colleagues here on the call had said is that looking at kind of the, the foreclosure, or the sorry the eviction process probably pertains to the foreclosure process as well. You know, it seems like it based on data it's a point in time kind of event, but it is a process that drags out over time and there's various benchmarks. And all of that the impact it has on individuals and their mental health starts with the termination of tenancy, the stress and anxiety that goes with it, then to the filing of the unlawful detainer and of course this will be different across states. And then these different points in the process that people have to be able to understand and navigate and don't necessarily have the kind of legal support they need to defend themselves. So seeing it where those kind of pinch points are in the in the process where community members kind of give up so that we can see, you know, they're just releasing their rights because they don't know what to do. Thanks. So I'll just ask one more question of the panelists and then we'll move into a Q&A. So for members of the audience. If you haven't dropped a question in yet please go ahead and do so. So the question is this. We know that there is some discussion at the federal level of creating a national eviction database and capacitating cities and counties to collect and analyze their own eviction data in order to, you know, both have locally available data but also to feed into a national database. In an ideal world, kind of, you know, looking at your experience with EPLL and the feed tool, what types of resources would you need to have, you know, at the city level in order to allow you to do this type of work kind of in a sustained fashion. Lauren Lowry, I'm sure that you could answer. Looking across the board of city is what you're seeing as the biggest. Yeah, just kicking it off. I can't emphasize funding right funding for staff as funding for capacity building and training, also incentivizing local collection standards I think that's important. Creating eviction data standards as well as providing technical assistance to local jurisdictions, but very much I would like to underline funding. I'll just say yes yes yes everything that Lauren said, and you know and I think that it's really helpful. Yes, we need a streamlined process nationwide but there also needs to be some ability to tweak to local needs right and allow for some additional innovation so you know striking that balance I think is going to be really important. Just again echoing Lauren and Sarah funding with an underline and an exclamation point and an all bold and italicized everything you can do in word. But I think it would also be helpful to have that technical assistance but specifically around how to talk to other funders and other partner agencies and our local elected officials in our state elected officials because we really need to have these conversations and get buy in at that level. So I think that would be an amazing resource. Yes, yes, yes, and yes. And I think I was going to say something similar to Lauren is that we really need to use this data to leverage policies that are at the state level that constrained the local level from providing the kind of resources that we need. The whole concept of evictions as a means to resolve disputes is not effective and has it makes people lose housing and it, it impedes their ability to find new safe and habitable housing. So looking at, you know, eviction data and what it's cost, and what it's cost to community members, and then looking at is there a way that we can do this dispute resolution process better, intervene earlier and address some of the problems so that it is not. People are not losing their housing over minor infractions over their lease agreement, or because they've gotten such a large rent increase they can no longer afford to pay rent. It's kind of fun. Common theme, and I know I said that I would turn it over to the Q amp a but I actually have just one last question that was on my list and I would really love all of you to answer just very quickly. We've had a lot of interest from other, not just cities and counties, but other partners, research universities advocates, you know, who are interested in picking up and using the feet tool. And I'm just curious, you know, if you had any advice to somebody who is interested in using this tool. What would it be. I think I would say this is a journey. Right. A journey that, as you said, a number of cities as well as states, we're on a continuum. Right. And so, there may be points and times where you have to take back and assess where you're currently at but that does not mean that the, the journey stops there. That means that you have to build relationships and partnerships and you have to begin strategically game planning, what needs to happen to go forward. Yeah, I'll add to that that I think the tool really allows an opportunity to bring partners together for a conversation about what, what data do we have and where are the blind spots, and to really start thinking about what the needs are and so you know I would think if someone who is interested in this start that process, because even if you wind up not utilizing the tool, there is benefit in just being able to have that check in and kind of know what what is known and and where the where the major blind spots and where is their opportunity for more informed decision making and you know, action planning. And I think that is why wildly beneficial above and beyond the output that the tool provides. My advice to any other localities who would use the tool is to put together a great team to help you through this journey. We had a team made up of you know I'm in the Community Development Department. We had our IT department, we had our city managers department social services, we had several nonprofits are local legal aid, and they've just all come together around this eviction issue and the data, putting together this data and it's just been very helpful to have this group of people at the table, who can again sustain this process after you know, the tool is done so. And I guess my only advice would be I agree with all of the other comments but we also have looking you look at data and it seems so far removed sometimes from the people. And so just making sure that as you, when you're trying to humanize the data in a way and how it impacts your community and who it impacts. It is a great way to tell a story, and to show what's going on in the community. So having that data is so important and while it may be challenges and you may have obstacles and you may have to really struggle to find the right door to get the data. It is worth it in order to be able to tell the story and to identify what the community needs are. Fantastic. Well, thank you to our fantastic panelists. We'll just have about 15 minutes for Q&A and Manu, Mallory and Sabiha if you'd like to come back on camera that would be fantastic because several of these questions are to you as well. So I'll start Sabiha with a question to you about how the partner sites were selected. Sure. Yeah. Yeah, thanks for the question and so the partner sites were selected in large part through our informal partnership with National League of Cities and Stanford Legal Design Lab who are co-facilitating the Eviction Prevention Learning Lab. So I think the majority of those 14 partner sites are also part and the three cities represented here today are all part of the Eviction Prevention Learning Lab. So given this, you know, the sort of the timing of these projects align such that Eviction Prevention Learning Lab was gathering 30 cities from across the country who are, you know, invested in using data and other sort of resources to really save off evictions at the local level. And so that's I would say the main mechanism for how the partner sites were selected and it really was not necessarily a competitive process or anything. It was really having conversations with cities who were interested in helping us test this and just ensuring that there was alignment as it related to data access, data, you know, technical capacity and just capacity in general to work with us. And then similarly for the cities that are not part of EPLL, most of them are cities that we've partnered with or counties that we've partnered with in the past through our prior data kind of new Americas prior work on eviction foreclosure analysis. And I think someone also asked who all who are all the partner cities and we actually have a link on our website. So some of those what some of those links to the tool will take you to our broader eviction and foreclosure data work and there's, there's more information even a map that shows our partner sites, not just as part of this, the development of feet but even beyond that so you can check them all out there. Thanks to be here. Next question is to Manu and the question is around geocoding. So, you know, what is the geocoder that was used as part of the tool. If you want to say anything more about kind of the geocoding capability is and how you used outputs of the tool to produce maps. Yeah, sure. So there's I mean the bulk of the geocoding work here in the tool is done by the census back geocoder API which basically means that you can feed in a batch of addresses, street level addresses into that API call and then get back all the results on, you know what the census track ID for that address is. It's a public resource and then so you can essentially, then go to the census website. I mean they also have a kind of interactive version of that geocoder that you can use if you want one off address and like which one do it one at a time. You can also throw in files into that like if you really want to do it manually and not use the tool you can actually throw in files up to I think 1000 or 10,000 records and it'll get get you back a file. And so it's against a public resource and go to that. And we can do it. The other part of the geocoding here is a hard resource or housing development resource which is a zip to census tract mapping. Sometimes like you know that the census geocoder is not like 100% accurate in the sense I doesn't always come back with an answer for you in terms of geocoding typically are accuracy is about 90%. And so 10% of the addresses are just not going to get geocoded no matter what you do, even if you clean them very well, it's going to get about 90%. And so the other HUD one, it's, there's no exact mapping it's a probabilistic mapping it tells you okay, like this is a zip code a zip code has a lot of different census tracts right so the way the tool works is it says okay well there's maybe five census tracts in the zip code. So let's just take the one which contains the Alice like randomly take one in like proportional probability to how much of the addresses that contains right so if one of the census tracts is half the, half of the addresses and that as a code will take like half the chance that you get that census tract and half the chance to get something else right so that's basically how the second part of it works and the way we do it is like and we know it's not 100% accurate but it's going to do at least geocode something rather than throw out data right and so that's that was the approach we took. Thanks so much Manu. So this question. I'm maybe going to start with Mallory, but I'd like to actually throw it out to our panelists as well and the question is from a county where eviction data is not digitized and it's not accessible, unless scan. The question is, you know, how can places like this county use the tool, if at all, and sort of the second part of the question that I think it would be really helpful for our panelists to chime in on if you have any experience is, you know, how do you move from a place of not digitized eviction records towards digitization. Julia, this is a question that we actually encountered multiple times in our earlier partnerships with you when we're trying to access eviction and foreclosure data from cities and counties to work on our displace in America and displace in the Sunbelt report. This is a challenge that many cities and counties across United States face. Right now, in using the feet tool, we do request, you know, the data to be digitized, such that it can be then formatted into the data format that is required for the feet tool. However, this feet tool is really just version one of the tool. And from the conversations today and the questions that the audience members have been asking, we're really starting to think about new ways in which we can enhance or add new modules to the tool. So there will be the potential for us to explore something like optical character recognition, which allows us to really lift from these non digitized pieces of paper, the information that is required to then format the data and it put into the tool. This is definitely something that we're learning a challenge that we're looking to see whether data science can really help unblock so that we can really enhance and increase the users of the feet tool across the country. But I'd also be, you know, really interested in hearing from cities and counties perspective as well around the steps that they could take to to also overcome this. But to add, it may be good to look at the city of Boston, because they ran into this problem, and they, they took a time to actually get their data up to standard, and you can find that at the city of Boston dot gov where they talked about the process that they went to to really digitize their addiction data. I'll just, I can also just jump in and say, you know, I think there's two sort of parts to that question. One is sort of everything that everyone talked about and their answers to that last question on the panel discussion which it, you know, investing in creating a digitized version of this data is very time consuming but also worthwhile if that is the only way forward. And so I mentioned, you know, hopefully there will be tools available to make that easier. And then the other thing I would mention is I'm not, I'm not aware of the specifics around this county or city. But, you know, even if that is the case, the data is likely digitized somewhere. And so oftentimes, even if that's what's being shared publicly, there might still be avenues to, you know, as Christina mentioned, figure out who is the person who has access to digitized data, whether that's the court system or a PIO, a public information officer or someone else. It's worthwhile doing that investigation to figure out whether you can kind of go up the chain, or around what's publicly available to. Great. Thank you. So we have time for, I think, just one more question. And, yeah, I think that I'll throw this out to anyone on the panel for anyone of the participants who's interested in answering. Both Lauren Lowry and Lauren White talked about how to catch evictions early. So kind of catch people who are at risk of eviction potentially before evictions are filed in court, right, because once you have an eviction on your record, even if that eviction is ultimately dismissed, still under record. So any thoughts that panelists may have on how the tool specifically or just eviction data generally can be used to sort of get ahead and get upstream of evictions. And, you know, as an early intervention tool for people who may be vulnerable to being evicted. I would say that the attaching the demographic information to the eviction record show us information about the income of individuals, who's below the poverty level, that specific data that really kind of stresses is this an economic issue, something where community members just don't have the capacity to pay their rent. There's a lot of when we talk about evictions and we are looking at the narrative from both sides of it landlord and tenant, there's a lot of narrative about, you know, bad tenants get evicted and the reality is when we look at the demographics there. It's hard working people struggling to make their rent. And so when we know that it's an economic issue and they need financial support, then we have to look at programs that would help either one provide that financial assistance or support the tenant with mediation to be able to talk to their landlord about the issues. And I think in a lot of complaints that we hear from both tenants and landlords there is a disconnect between the two parties and having somebody to be able to mediate conversations. So that they don't escalate so that they're able to address the problems instead of taking it personal. So figuring out who is being affected and most impacted by evictions then helps us to frame the solutions. And I would just add the historical timeframe to see layered on top of the demographic data to really give you an idea of where and when it's happening and how prevalent is happening in a particular location as well. Yeah, I completely agree and was just going to add that the question sort of, it sort of gets at that the data that is being ingested most likely into the tool is formal eviction data and so anything that exists outside of the formal court process. Exactly what Christina and Lauren said it's really identifying areas historically where those eviction rates and foreclosure rates have been high and then targeting those as a proxy really. Well, with that, we are at the end of our time. I would like to thank all of our panelists and participants for taking the time to share your experiences with us. I would like to end by saying that we are actively continuing to develop this tool looking through new functionalities and also sharing it with new audiences. We're actively doing demos and run-throughs for a lot of different organizations who are interested in using the tool for a variety of different use cases and that includes not just city and county leaders but also journalists, researchers, advocates. So, if this does seem like a tool that would be useful in your work, please don't hesitate to reach out to us. We would love to share it with you and walk you through how to use it. And with that, I will thank everyone once again and wish everyone a lovely afternoon.