Million Neighborhoods: Mapping Slums on a Global Scale

hi everyone my name is Syd a German I'm a graduate student in computer science in public policy at the University of Chicago hi thanks for coming I'm Cooper neder hood I just graduated a master's program from the University of Chicago and I'm currently working for the manse way to Institute what's the man's way to incent you you ask well it's a new institution at the University of Chicago dedicated to studying the forces and processes that shape life in urban centers so questions that researchers at the Institute look at are like one of the effects of having green spaces as you walk to work or when people are displaced because the city is being renewed where in the country do they go do they stay in the same city do they move around the country and one project that we particularly work on is figuring out where across the world under serviced neighborhoods also sometimes known as slums are actually located so that's the million neighborhoods project and that's what we're here to talk about today all right so we're interested in mapping informal settlements and you might be asking yourself what are informal settlements and I think something you know this picture right here is really as illustrative of the problem we're trying to look at here so in the back you can see these really tall big urban buildings and that's kind of what we would characterize as quote/unquote a formal formal urban area and then in the foreground we have these kind of smaller really clustered area and that's what we would characterize as an informal settlement and we're really interested in looking at how these built the way we build our physical environments affects the lives that everyone can live in all of these places so just a really quick agenda so we're gonna talk quickly about the project background kind of some of the theoretical and research roots that this is coming out of what our task was and then really all this culminates and a global analysis where we're gonna try and identify all of these sort of places that all of these informal settlements alright so why do we want to do this you might not notice this but the world is urbanizing currently extremely quickly and extremely rapidly people are moving into cities very quickly and currently 4 billion people and a lot of this is happening in developing countries and it's often kind of unmanaged and kind of haphazard and you can which results in some of these informal settlements that we've that we were just looking at and they're often characterized by a lack of access to public infrastructure whether that's clean water sewage and so characterizing those informal spaces is important in ensuring that everyone can gain the full benefits of this urbanization process that is currently undergoing all right so this is coming out of a bunch of work at the University of Chicago kind of two main contributions so first there's weave and past past work kind of authors listed here have demonstrated that it's possible to measure the lack of street access using a topological analysis so we can get more into this in the appendix but looking at the connections within these neighborhoods and - once we've used this to identify areas that are that are lacking in access we've we've developed a method to build a street network which can then deliver universal access to all of these previously enfranchised inhabitants and the reason we focus on street networks as our proxy for services it comes from conversations we've had with activists and residents of underserved neighborhoods who point out that the peripheral Road around an informal settlement is actually where most of the formal infrastructure whether it's drainage sanitation clean water electricity or even Commerce that's all basically connected to the existing road network so that's why that's the central focus of our approach yeah that's a grit that's a great point so just kind of back to this picture likes attention was saying often the EM the access to public infrastructure is around this kind of outer part and people in the middle of these dense areas are often kind of without access to public infrastructure all right so our task is to scale up this initial research and to apply it globally and the first part you know how are we going to how are we going to do this first we relied on OpenStreetMaps as a global open source data and then we had to apply this general technique across this gigantic data set and now I'm gonna turn it over to Satish who will introduce the website so this is actually our website that we built this is the results of all of our analysis I I picked the color scheme which i think is the most important part but basically every distinct color you see here is a delineate of street block so there's a single boundary road around this and there are multiple buildings inside inside this street block so anytime you see a blue block here that means that it has relatively good access so from within each built from within each building you have basically direct access to this road network around you the blocks in red are what we might call impacted access because you have to cross through other people's parcels of lands to get to to get to these formal services so this is a zoom in of port-au-prince in Haiti and if we zoom in a little more you can see that some of these across the city there isn't just like one characterization of these blocks like in this same city you have some very regular blocks up to the north and some very irregular blocks to the south so in general what we found is that these well-connected blocks have better sanitation power all the services that we've been talking about and as we zoom in a little more we can see every black dot here or every black polygon is a building footprint that we pulled from OpenStreetMaps so it was a little hard to see from like the zoomed out view but as you zoom in further and further to some of these red blocks you actually understand what the density is like so if you say you live in a settlement over here where my mouse is moving around and you need to get to I don't know an ATM over here you have to cross through a lot of basically unexplored and unprotected and unserviceable it's jump quickly to Nairobi in Kenya and Nairobi is is pretty interesting because it has a well-developed central business district so these are the central areas in blue here so again kind of the same store we saw in Haiti that there's some blue there's some red but what's really interesting about Nairobi is just like beyond this like well connected grid system where you can see that basically every building here has proximal access to the street we also see within the bounds of the city the Africa is the largest urban slum and that's called Kabira so just a couple miles from that central business district it's it's really not far away so like basically everyone who works in the central business district around here in the top right plays golf in this circle right here the Golf Course and then right next to that Golf Course is Africa's largest urban slum so this is very much within the heart of the city and you can see again what the density looks like and the fact that it's hard to run up services like power and sanitation to each and every building in this era area so with our approach of using roads as a proxy we realize that there's a really standard computer science formulation of solving this problem so we figured that out that there was a way to connect every single building in these settlements to some sort of Road and figure out what the minimum cost road is whether that's measured in terms of length measured in terms of disruption or whatever the local community's needs are and for a number of neighborhoods we've proposed these new roads highlighted in green too that would show what it would take to actually provide services to each and every single settlement here Cooper do you want to talk more about that the algorithm there so getting in the weeds here a little bit but essentially it's a slight generalization of a minimum spanning tree but and we're happy to answer more questions about the details of implementing it and it was definitely computationally intensive to just do it over such a huge huge craft here but the takeaway really is that you know there's is that this is kind of the the minimum cost in a sense Street network that would you know give all of these internal parcels access theoretically to public infrastructure yeah so our hope is that this isn't a top-down kind of dictum about how you need to build your communities like what we realized is that these are really just seeds for plans that local communities can take and realize like hey this actually makes sense to go ahead and invest in these roads versus no this doesn't make sense we don't want to build a ton of tiny roads like these people are willing to move and we don't want to invest in a much larger Road those are the kinds of conversations that we've had with activists and with NGOs such as slum-dwellers International and even formal governments such as like the the mayor of Freetown has been very interested in this work I mean she comes from a background of slum upgrading so she really is in with figuring out like what's possible and not possible but we definitely want this to be more of a springboard rather than a final say of how these communities could upgrade themselves and what Institute in situ slum upgrading could look like so yeah like right now we have a good chunk of Africa analyzed in terms of where which neighborhoods have access and don't have access and in the next couple of weeks we're planning to push updates to Latin America where we have some of the some of it analyzed and by February we should have the rest of Asia filled in so this is all live on milling neighborhoods dorg happy to turn it over to any questions and we'd love for you to check it out and let us know if you have any thoughts this is a yeah I think that's an absolutely valid question and like surveys of residents of these informal settlements definitely ranks 10th your security and eviction is one of their fears what we found looking at these like historical cases of eviction is that the government's pretty much already know where the slums they want to evict are like they're they're being targeted because they're like next to you like a golf course so the government wants to turn them into like high-rise development because like the golfers don't want to see the salaah was won by the 17 pool or something like that but that being said I think that that's definitely a valid concern and our way around it is to just be in constant contact with people who live in these communities and see like what what their concerns are and for the most part people are in these communities tend to see the benefits of knowing what a possible upgrading solution are outweighing some of these privacy risks but that's a great point studied like this around the University of Chicago in the area yeah absolutely it's actually sorry go ahead yeah so the university itself is kind of interesting because it shows up as a false positive like when you're wandering around the university you can see like there there aren't road accesses too like in your buildings within the university so our algorithms like oh yeah the University of Chicago that looks like a slum and depending on your ideological orientation might agree but in Woodlawn itself based on its history I think it actually shows up pretty well it like most of the the buildings have some sort of Street access to the to the formal road network it's not the case where you have a too many blocks with like informal structures that impede each other's access to the infrastructure so and I think that's just a legacy of like how it was built in the thirties and forties yeah and kind of to that same point I mean it's important to kind of keep in mind you know where our methodology fails and that's so we had this slide here you know in some of the places that we've looked at the it is you know a more of a question of accessibility but we have there are plenty of neighborhoods like you know false negatives where this is a community in in Mexico City where it has a very formal physical environment but it still has lots of urban challenges and you know it's important to recognize not only where our methodology works but also when it doesn't when it doesn't work and I think there's still something to be gleaned you know even in these false false positives like potentially the campus of the University of Chicago and also some false negatives and kind of what that can tell us [Music] my question is do you have a strategy to go with this of teaching people to tell stories using the data so that can make sense of it in different local populations because yeah I think that's a great question and that's something we were just having a meeting today earlier today actually kind of trying to you know cuz there's this huge map it covers the whole world but kind of synthesizing some of that information and operationalizing it so that it's useful to researchers whether it's other researchers people on the ground I think it's something we're actively thinking about and we're you know part of this process is to is to then connect this this information that we've generated with other people on the ground and you know see how this can be useful to them how how would they like to see it develop LexA - so Paige mentioned this is you know really just we're hoping that this is just a first step and we're definitely hoping to build this out and make it more and more useful to people on the ground yeah that's a great question I think a lot of the work was done in Freetown in Sierra Leone as well as Monrovia and then I think there was additional ethnographic work that some of the earlier researchers on this project sampled based on Harare in Zimbabwe as well as neighborhoods in Uganda so it's generally been sub-saharan Africa with a handful of inputs from some larger slums in western India such as Parveen by and Foye body in Pune crowd source is other areas where some of that data is out of data underserved yes to both of those but let me see if we can actually yeah so data completeness is definitely something that we struggle with because essentially we're in a lot of places either overestimating or underestimating the complexity if we have more data about building footprints versus roads or vice versa and based on like composite data that we've pulled together from privates satellite imagery providers and some government and non-government organizations we think that in a typical country like Sierra Leone our building for footprints are about 50 percent complete and that varies around the region and I think there's probably a similar issue with roads in that yeah like that there just aren't mapped out roads despite a couple advances in using computer vision techniques to automatically add them to Open Street Map despite that one strength of the Open Street Map is that a lot of slum upgrading and crowd-sourced geo referencing uses Open Street Map as like the engine to coordinate where the data is so a really interesting case is Lesotho where the government has decided that instead of hosting their own infrastructure for determining we're building footprints are they're just going to use Open Street Map is like the official repository so like the best source of data for infrastructure in Lesotho is actually Open Street Map which we didn't realize until processing Lesotho took three times as long as any other country in Africa so then we started looking into that and found all this information about how the government had actually funded the use of Open Street Map as like the official database for urban infrastructure in Lesotho so definitely concerned we're always looking for new sources of data and we're actually data source agnostic so if you have some hot tip about data let us know good evening my name is from hailey probably for some of you this data family means something but it doesn't mean that much to me means a whole lot I wish I could do it myself and do something with it with that being said thank you very much for your work your studies your research the time you put into it I have seen kids walk in about three hours my feet with gallons of water meanwhile try to go to school crossing rivers Daejeon because sometimes they don't know how deep this river is I have seen farmers young women telling the basket of fruit in the head walking two hours three hours to put try to study to make money to survive so I was doing 80 in 2016 I was like I left a baby when I was 14 I think something was changed but it gets was right after the earthquake things get worse this is why I have a master's in theology I decided to go back to school get a masters and project management now they are young people with envy and a lot of people who know things who live outside of the country there's the government everybody's intelligent but not at the same level right so I'm just being political right here so data is like that we do this studies I know what should go down but sometimes just rope soft and a corner because it is very difficult to get a quick talk to do what we need to do with it now the situation is pretty bad and I wish that this is not going to be a waste of time because sometimes you study things over here and you wonder why I'm gonna do with that okay next time you know I'm working in a warehouse right so this is very very important but wasteful thank you i we we don't know when we're we're exactly the breaker is but we don't want this to be just like a theoretical academic project so we'd love to talk more about you where you think we could actually make the rubber hit the road with this data so I work for you're particularly with socks so adult so you know you know those are and when you talk about this is a starting off point a jump off point for constructing it you know construction about new land new information to be that sounds like you are you're hoping that governments create kind of business out of the data to provide a with that so my question to you is what kind of outcomes may be seen for providing this data to local governments with some expectations do you want gums to maintain data and to see what progression or lack of progression has happened and I personally think that's a part of survey methodology but it could be something guys I think an ideal outcome that we'd like to see I think we envision some of these things as you know like our last gentleman said we hope that this can be can make the process of Neighborhood Development easier and give people that are because sometimes when you're trying to you know we really I think are ideal you know outcome would be for this to be a tool that communities can use to help inform you know their process of developing their own community and I think because there's you know the process of organizing all these disparate opinions and getting people on the same page you know that's kind of the the organization - that can be difficult so I think you know we really hope that this can be something that it's easier to give people a menu of different options and for them to you know debate amongst those and rather than just kind of you know ex nihilo yeah I agree with Cooper and I think the the handful of like nascent efforts in this field are kind of like government intervention agnostic it's basically groups realizing that they actually have enough capital saved up in the community figuring out ways to finance the construction of some of these reblocked roads and figuring out what it means for their area to actually improve their infrastructure without the government actually financing or even directing it that being said a handful of like city and local governments are interested in knowing like where how their cities are changing because a lot of places in the developing world the last formal census was done decades ago so trying to figure out actually where populations are concentrating and where services are needed most is one thing that local governments are super interested in using this data to figure out so basically the problem is that is coordinated with the actual nation it's you know they do both mostly city-states you know if the structures interested the actual or target a country name that's right yeah thank you for your presentation it was very informative I wonder where did you get your elaborations from moving forward where do you see collaborations needed I think the the genesis of the million neighbourhoods project really does come from a non-government organization called slum dwellers International and it's basically a network of community leaders throughout the developing world trying to figure out what works and what doesn't work in improving the quality of life in these informal settlements so I think a lot of what we like yeah a lot of what we think of when we think of informal settlements really does come from that area of work like I think everything that we do is pretty much aligned with their theories of change and what we'd like to do is kind of bridge the gap between this like grassroots bottom-up method of organizing and try to connect that to infrastructure and capital whether it's figuring out new ways to finance a figuring out better data sources through like formal government or non non governmental means I think bridging those two worlds is what is something we hope that this tool is able to do [Applause] you [Applause]

Channel: Chi Hack Night

Views: 194

Rating: 5 out of 5

Keywords: civic tech, civictech, opengov, Chicago

Id: QU-NhpsfhRA

Length: 27min 34sec (1654 seconds)

Published: Wed Jan 15 2020