 Good afternoon friends, on behalf of the Center for Urban Science and Engineering as well as Indian Institute of Technology Bombay, I have immense pleasure in welcoming Professor Eric Isaac as circulated in the email, Professor Isaac is currently the provost of the University of Chicago. He is basically a professor of physics and formerly director of the Agrone National Laboratory in the US. Now he will today talk to us on solving the grand challenges of the world's cities, urban science and energy and specifically he will focus on how urban challenges such as pollution and traffic can be addressed in partnership between data scientists and the policy makers. We have some data scientists here and some of us are in the policy making. So it is going to be a good learning experience for us. So basically using the search standard of the energy policy institute of the University of Chicago. So on behalf of the institute we welcome you and so glad to have you here. Thank you. Thank you and I really appreciate you hosting me Professor Narayanan. So I'm Eric Isaacs has just heard and my title is, I took out the word grand. It's a grand title basically trying to solve this urban challenge and so what I think I'd like to do is actually ask a somewhat different question is what can we in universities do about these urban, these grand urban challenges. And so let me start first by just explaining a bit about the University of Chicago. Some of you may know us. This is a picture of us. This is downtown Chicago. So we're just south of the city. Some of you may have visited us in one time or another. And the reason I think universities are really well tooled to do these kinds of major challenges is because the urban challenge of course as we've already heard is not just data scientists or just a bunch of physicists like myself, a bunch of geeks sitting in a room. It really is a true collaboration. It's scientists, it's politicians, it's people in non-governmental organizations all working in some partnership because even as academics as we are you have an answer. It does nothing to blow, just to holler into the wind. You really need to think big. So a university like IIT, Mumbai has a lot of different types of disciplines. And so at the University of Chicago just by way of talking about urban we have five divisions. We're broken into five divisions, physical sciences, humanities, social sciences, a college, which is really where we teach our undergraduates, and a biological sciences. So if you think about urban challenges you've got the technological challenges, so you need physical sciences, you've got certainly health issues, so our biological sciences is a hospital but we also have biologists, think about biology, think about disease. You've got real big social issues and so in fact social sciences we visited yesterday with the Tata Institute of Social Science which was originally, the original director of that was the University of Chicago faculty member who came from what we call social services administration, part of social sciences which is all about social welfare. When you think about big urban cities of course social issues are tantamount. You can't just establish, collect data, say well we're going to do this and this and this, you really have to understand people and understand how people behave. And last but not least humanities, you really have to understand more than just the science, you have to understand the humanistic aspects to it. So a university really is well suited to think about these problems. In addition at the University of Chicago, we have, if I can get this to work, I may just do it by hand, we have six professional schools. So we have the Booth School of Business also thinking about urban problems. Even if you had a solution you've got to scale it, you've got to think about how to make it something which is business worthy. We have Harris School of Public Policy, policy is very, how do you write good policy that actually can be followed and some of the things I'll talk about are data driven policy decisions. The Harris School is a critical part of that. SSA I mentioned social services, these are actual social workers. They do interventions and families to think about the social issues. We have a law school thinking about writing laws. So when you think about the urban challenge, all of these will play a role. And a medical school of course, health. And above all of it, the University of Chicago, like IIT, Mumbai is embedded in a major city, a really important city in the US. Mumbai is, one could argue, the city in India. Maybe if you're in Delhi you may not agree. But we are actually located on the south side of Chicago, which is where a lot of poverty exists. So when you think about affecting society, you think about, we live it every day in a sense. Our medical center takes 40% of our admits are from the south side. Lower income, poor people. So thinking about the problem holistically is extremely important. And one of the things that characterizes the University of Chicago is the fact that we have many institutes that are designed to draw together the different components of these divisions. And the one you'll hear me talk about in the next half hour, 45 minutes, is the Energy Policy Institute at Chicago, which thinks about, an economist's point of view, Michael Greenstone, we just hired from MIT, thinks about the problem of urban environments, thinks about the problem of energy and sustainability and pollution in the context of an economist in terms of not just policy, but also incentives and et cetera. So I thought I would just give a little background. Probably most of you know a lot of this stuff. But this is the per capita energy use. Just talk about energy. I'm going to talk about pollution or talk about carbon. Energy is per capita versus GDP, gross domestic product per capita. And this is a chart that was made from the International Energy Agency's data by someone at UC Berkeley. And, you know, by hand. So of course the USA, other OACD. But the ones I want to highlight, of course, are India and China, which are poised to explode, already in a position where things are just changing extremely rapidly. And the question really here is what? This is just a, if you will, one of the data, which maybe some of you have even seen, which came from the World Bank, which just is a, if you will, it's a spectator or it's a indicator of the kind of growth you're seeing here in India. This is just a number of air conditioning units. It's not necessarily the only thing which talks about energy use. But if you think about electricity, it's basically the use of electricity is just essentially linear with air conditioner use. And this is a good thing, the fact that it's a healthy thing, it's a positive thing. But this kind of growth, which is 14% annually, and I think it's actually even accelerating in the last year or two. This data is already about three years old, is remarkable. And it's asking, it's begging the question, how are we going to handle the kind of growth here? And I'll get to Urban in a minute. This is also another chart from the International Energy Agency. It's their outlook, 2011, but it hasn't changed. I apologize, it could have gotten 2014. But what you see here is that if you look at the OECD countries, these are the economically developed countries, US and Europe, et cetera. And you look at the blue and energy consumption. In the next 10 or 20 years, it goes up a bit, but it's not skyrocketing. The biggest growth by far in energy consumption will be basically India and China. No surprise here, right? And you can just, it's very easy to see that. So that's just pulling out Asia. This is, with the increase in energy use, of course, with some models, with the increase in energy use comes an increase in carbon. So what you're seeing also is that these bright green, this is, again, China and India, a dramatic increase in carbon. If we essentially do business as usual, that means that growth is in China, it's essentially a coal plant being built, several coal plants being built per week in China, similar growth here in India. So if you think about that energy growth, you're also driving a large amount of carbon and other types of pollutants. And so this is the issue that we're facing. And so the question comes back to this one, which is cities. Are they the problem or the solution? So cities are really amazing places. Tremendous amount of population density, of course, drives not just negative things, but positive things. It drives innovation, it drives creativity, it drives an energy level which actually produces economic growth. It does many things. But, of course, cities can also lead to, by the way, this is a picture of Bangalore, I don't know if any of you recognize the city, but sorry. Chinatown in Bangalore? But it is Bangalore. Yeah, so, and here's the data, right? So this is actually United Nations data, which is to say, we've actually crossed in the last five or 10 years, we crossed a really critical point, where historically the growth, there was a larger number of people living in urban settings, sorry, in rural settings than urban settings. So what happened in around 2010 was there was this crossover to we actually now have more people living in cities than living in urban. So that means that more than 50% of the world now lives in urban centers. And you can see the projections are just going to continue to grow. China is driving this intentionally as an economic driver. It's happening here in India, partially intentionally, partially just because that's where the jobs are. So I just put together some data on the challenge here in India, and I'll come back to some of these. If you just look at India, I'll come back to some of these numbers as I said, but if you just look at India, it's about 250 to 300 million by 2025. The growth in cities will be 250 to 300 million, something like that. In China, the number is more like 400 million. So the growth in cities is just remarkable. And of course, the question you have to ask is, can we do this in a way which not only supports that growth, but is sustainable? Is sustainable from an energy growth point of view? And then, of course, when you grow energy, are you also doing it in a way which is sane, which is a minimal carbon, or at least some carbon. And here are some numbers. So right now, of course, these are numbers you guys know. I just thought I'd put them up here. 300 million people in this country lack electricity, period. So they're off the grid. And this number I'll come back to, too. 3.2 years increase in average life expectancy for 660 million. This is a challenge. It turns out you can do numbers. And I'll show you numbers from China. You can actually connect the amount of carbon in the air to lifespan. And if you start connecting those two, at least you know the information and whether that's an incentive for us to think about reducing carbon is a good question. And the other thing, which is not a happy statement, which is that 13 of the world's top 20 most air polluted cities are here in India. And I believe Delhi is right now the worst. It's actually gotten worse than Beijing simply because of growth and consumption. It's continuing to increase. So these are really big challenges. And the question or the vision might be to say, cities, can they be made livable through some kind of evidence-based policy, intelligent, energy-efficient, renewable technology, some combination of policy, technology, et cetera? So I'm not going to stand up here and say, I have an answer for that. But I will show you some of the things that at least at Chicago were starting to do to think about some of these big challenges. I've already said this. I mean, this is an interesting, just for visual, if you need a visual to really believe this growth, this is the Pearl River Delta in Guangdong province in China from 1980 to 2005. And it's just, and that's 25 years for that kind of development is shockingly fast. And you can just see visually what's going on here. And these are the numbers. 70% of Chinese people live in cities with 1 million or more people by 2025. 10 years by 2030, 221 cities will have over 1 million people. And this, I probably should have put in a major building here in Mumbai. This is Empire State Building in New York City. I'm from New York, so this was, but if you think about it, what China has to do is construct one New York City every year for several decades. They also have to produce one coal plant a day. And we can argue about throwing in nuclear and a bigger nuclear power plant. It's even more atrocious because you can't build nuclear power plants that fast. So these kinds of numbers are staggering. And for India, you can argue they're similar. I really apologize about this, but as I said, probably a lot of you know some of the numbers. I've already mentioned that 300 million people are without electricity. We'd like to turn them onto electricity, both from an educational point of view, a pure quality of life point of view, water, huge problem, and on and on and on. This is actually a chart that was prepared by IBM who's thinking carefully about smart cities, about the changes, and how to think about starting to manage these flows of people into major cities. So basic theme of this talk is what can cities do and what can they do with real information? And one of the things we always pay attention to it, University of Chicago, I'm sure you do here at IIT, is the data and understanding first and foremost what the data is telling us. We can predict into the future all we want, but unless we understand the data today, we don't have anything. So what kinds of things might you think about using data for? And I'll give you examples. I'll highlight some of these as we go through the talk. But if you think about data, what can you do with data? Think about policies and infrastructure investment. If you had good data, for example, on health, crime, I'll show you some examples in crime where we're having good data on different types of policies or experiments in education actually impact the level of crime and recidivism in the city of Chicago. Energy, of course, and low income energy, cheap energy. Road pricing, one of the things a lot of cities and maybe Mumbai will consider it one day is how we think about do we want to, London does this now, there's a premium for driving a car into the city based on what data. And the question is even if you start charging, how effective is that? And so asking questions like that, understanding what data to collect and then congestion charging. Other things you can think about doing is operations. And some people here are operations specialists thinking about traffic flow. I'll show you an example of an experiment. We haven't tried yet. We're about to try in Chicago to study all kinds of traffic, human and automobile traffic in the city to understand flow and questions. Can you do anything about it? I'll also show you a cool experiment done by Berkeley and MIT in the Berkeley area, in the Stanford area, which shows how more managed rush hour traffic can make a big difference in commute times, and et cetera, et cetera. So I won't go through all these. Infrastructure planning, increased regulatory compliance. And this I will give examples of that even if you have regulation, for example, on factories, is it being enforced and is it being incentivized in the right way? And I'll come back and talk about that in some experiments. And then the connection between, for example, environment and the impact of environment on things like asthma, on health. And if you can monitor things like carbon dioxide or ozone, can you then make connections to occurrences or outbreaks of things like asthma or other diseases? Of course, the answer to all these must be yes, but the question is how to do all these things and think about it. And I won't finish all these, okay? So the real big vision in all this, and this is where us scientists, engineers and even social scientists believe in data, maybe sometimes too much, is in principle, if you really do the right things, you make some really good models, you could move from what we do now, which is very heuristic, it's very random and arbitrary, which we were just talking about. Cities have grown the way they are in a very ad hoc. I don't think there's many cities that are planned. And could you actually go from that into something which is really based on clear information and data? And I don't know the answer to that. In fact, maybe none of us would wanna live in a city that's over planned, but considering the number of cities that are being built here in India and the number in China, it's a good opportunity to ask the question, can we do something? So we wanna go from a reactive to proactive. And one thing I wanna highlight, which I've already said is one of the keys, I think, to success in anything that you want to do in the urban space is we academics tend to have great ideas and publish and then say, I have an idea and throw it over the fence. You guys figure it out. It just doesn't work because most of the things that we think will work, we may have the right answer, but government will never succumb to our ideas. Really, you need to, right at the beginning, have clear relationships with government, non-government organizations and universities. So I think this is a huge opportunity for all of us. So what I'd like to do in the next few minutes is just give some examples of the kinds of things that we're working on in at the University of Chicago, some of them in the city of Chicago, some of them more global, but also I'll throw in some examples that I just like that have been done elsewhere that I think are fantastic. So the first one really connects, it's hard to find data which is so clear on things like pollution. You can measure pollution, you can measure changes in pollution, but this is a very interesting story that was done by Michael Greenstone, who runs this Energy Policy Institute. At Chicago, looking at the effect of life expectancy, the effect of pollution on life expectancy. So it's rare you get such a clean experiment. In China, they had a policy called the Huai River Policy. And that policy, in a way it was a terrible policy, maybe, but it was great because it was a great scientific experiment and so what they did here, this is the Huai River, this brown line running right down the middle of China and the policy simply stated was that in the winter or any time of the year, cold times of the month, not any time, the policy was that any residents north of the Huai River would get major subsidies for coal and most of the residents burned some form of coal to heat their homes and anywhere south of the Huai River, there were no subsidies. So you could pay for it, but what it meant is that literally you'd cross a bridge and you'd go from a house that was nice and toasty warm to a house that was actually cold because of course the river being, I don't know, 100 meters across, the temperature is pretty much the same on both sides. So they set this policy, which is an interesting way to think, so what Michael Greenstone and his colleagues were able to do, this is, he did this back when he was at MIT, was effectively measure, the first thing that is, they measured particulate content in the atmosphere both south and north of the Huai River as a result of a consequence of this policy and so this is degrees north of the Huai River, this is, zero is the Huai River, 10 degrees north, so this is south, this is north and basically measured the amount of total suspended particulate in just milligrams per cubic meter. Just a simple measurement really, it's done pretty easily and what you can see is a very clear, very clear statistically significant jump, literally at the Huai River. I mean you could say it should spread a little bit but this is during active season and over time it does but you can actually see the fact that there's a substantial 30, 40% jump, statistically real jump in, it's about 250 microgram, micro cubic micrograms per meter cubed jump as you cross the river and so then the question becomes, what's the impact on life expectancy? So they did a longitudinal study, the policy I think was implemented in 19, I think it was in 1980 if I remember correctly and so he was able to cover a fairly long period of time and then looked at life expectancy and this is just age, age of individuals and this is the same horizontal axis and what you can see is also statistically significant, obviously the size of these circles has to do with the uncertainty in the measurement. You can see as you move away in degrees north or south, there's a jump right there and you can actually integrate under these curves and estimate the total impact on life expectancy but basically at the border there's a five year, almost a five year jump. So this was real statistical evidence that something was going on and so the thing in China was that he published this data and China hasn't done anything about it, the policy kind of came to an end but it was very interesting because it just showed what the impact of coal, burning coal was so what the estimate is of course is if you think if you just do all the numbers right and you do this it's a bit of a PR stunt at some level on the other hand if you just say that and you integrate those numbers, it's about 500 million people in Northern China are losing more than two and a half billion years meaning about five years of their life due to the pollution. So it's a rough estimate but it gives you a sense that the kind of pollution you're seeing actually does have an impact on life expectancy. Now in the case of China so it was real good data and it showed something. In case of China it didn't inform any policy not yet so maybe it will inform policy but in the case of China it basically didn't. So Michael Greenstone has just recently started to work in India in the last five years and has done similar estimates here in India just by looking at the impact of the China experiment which basically measured the particulate matter and if you measure particulate matter and do a rough almost a back of the envelope calculation we're talking about three years loss in average life expectancy in India. This of course isn't, this is a statistically significant number one that has to be looked at and appreciated for what it is but if you just take 3.2 years out of the life of Indian society it's a big number and so it's not so different from China and the question is does this suggest some kind of policy change? So it's data, here it is. This is an average over Indian terms of particulate amounts so this isn't published yet though. Oh yeah, yeah okay. And what were the numbers there? Three or four years, yeah. So this is remarkable, it's something when you quantify it this way I guess this is for scientists you say well that's just multiplying the number of people times life expectancy but if you talk to a politician about that kind of number it makes a big difference. If you can say you can shave, even if you can do a shave a year or give someone a year more of life it's a big deal. So this is a story about pollution associated with energy and the use of energy. I'm gonna change my gears a little bit here and talk a little bit about infrastructure. This is one of my favorite bridges. It's a beautiful bridge and I guess you can say it's helped alleviate traffic in Mumbai but it also hasn't solved the problem but it really is thought of as a way to alleviate traffic. So and I'm just using this as emblematic of what I wanna talk about but one of the things I'm thinking about this is the, those of you, I mean everyone here is probably driven on the Bhandar Wurlai ceiling. The thing about this, the only thing I'd say about Bhandar Wurlai is it does kind of drive you to think about automobiles and not other solutions. And so the question is, is this the right approach to alleviating traffic and that's just a question I'll leave hanging in the air but it's a fantastic bridge. As I understand it actually, it was supposed to extend all the way down the coast. Is that right? It was supposed to, it would have been much better. It would have taken a lot less time to get down to the, anyway. So I wanted to show you one experiment which I also thought was very interesting which was done by, which is related to traffic and infrastructure. Just to give you a sense of how data can tell you a lot of stuff about what could be simple policy decisions and I'll explain to you what a policy might be. There's no policy implemented from this data. What a policy might be. And this study was done by UC Berkeley MIT team a couple years back and this is the Silicon Valley area, Palo Alto around Stanford and Silicon Valley. I don't know how many of you have driven on 101. You know what it's like? Yeah, so you think twice about going out any time other than at the extreme. So they said, well, they were able to make a deal with the phone company to essentially be able to get decoded data. So in other words, just using GPS data. Not knowing who, you know, privacy issues so they didn't know who it was. And they did some calculations based on the GPS data from cell phones to think about how they might use that data to reduce congestion. And I won't go through the formula on this but what they found was they did some modeling, right? Delta T, so if you think about this is a big number here this relates to the total commute time. So what they did was first of all this is just a distribution of probability distribution of commute times TC in minutes. And you can see actually extends way out to many, many minutes. This is five hour commute times or more but this is not, by the way, single cars. This is total commute time. It turns out that actually the numbers are really impressive because I gotta remember what they are, but the total commute time in, I can't find it, but the total commute time if you integrate overall, okay, so if you look at this chart the total commute time is something like it's many billions of hours for San Franciscoites. It turns out that if you use the GPS data you can actually reduce about four billion hours from a total travel time and the reason you can do that is because I'll show you what that looks like in a minute. If you do that of course you make people a lot less stressed but more importantly that's a few billion gallons of fuel as well so it's a very important connection. So what they did is they said, well they did a good Duncan experiment basically. They said if we could, if these are curves so they did the Bay Area in Boston also have you driven in Boston. Boston is equally frustrating and I went to school in Boston, it's horrible and it's like the Bay Area. And what they basically said is they said, okay, if people drive randomly there's a measured result and so that's this green line here for the Bay Area and it's this line here for the Boston area. And what they said is if we pulled out, they basically said if we pulled out 1% of the drivers so we were able to determine, we were able to signal people using their phones or what have you, don't come out until such and such a time. Delay by as little as 15 minutes so meaning they would tell people when they could drive which is an issue in itself so it's not a policy that's necessarily implementable but they claim that if they could select 1% of the drivers, the right drivers but they could select 1% of the drivers they could reduce congestion by 20% which is huge. When you're talking about the many billions of hours that are consumed and driving and this basically chart shows it and there's actually a detailed calculation which goes with this but the point is that this is the 1% and they could actually do, these are dramatic changes in the ability to reduce the congestion. So it was an idea that hasn't yet been used for any policy but again it's interesting data that does something. Here's some other interesting data that hasn't been used yet. This was actually also MIT. I said I would be using other people's data. They had something called the Sensible City Lab and they did this study, this is Rome and what this is that red is essentially the cloud of GPS users so it's the telephones, everyone's telephone and these yellow dashes are, let's see it's a movie so I'll get it to move for you but it's these yellow streaks are buses. So the question they were asking was could they use the GPS data and the bus traffic to do something intelligent about how the buses were being shoveled around the city and the answer of course is yes. The buses here, I mean they're reasonably located where the people are but one could do in principle a better job at routing buses if you knew if you could collect that kind of data and so this data is always available. There's nothing fancy about this. You have GPS's on the buses, you have GPS on every cell phone user. As of yet, it's not being used to set policy but it's interesting data nonetheless. I thought I'd throw this in sort of the last slide on transportation. This is something you may even know about this project. This is the BART Rapid Transit project where it was actually very innovative in some sense and to me when I first looked at it someone showed it to me and said well that's obvious and simple but they actually put in this exceptionally extensive it's now about 50 kilometers of bus service through the city which you know so instead of encouraging cars they encouraged the bus service so it's a real simple solution. Easily accessible, connected up with the train system. I mean it sounds obvious but it actually transferred it did shift a lot of traffic, road traffic into buses. So it's just thinking simply is often the answer that you want. Okay so I wanna change the subject a little bit because I've been talking about data and the question is can this data influence policy? You know as a scientist I think well great science again you know come up with great data you know the data that MIT took in Rome so someone ought to do something about it and the question is is that science data and technology enough? And the answer of course is not. So I wanted to show you another experiment which was done here in India by Epic by the Energy Policy Institute at Chicago but really asking the question if you have good data how can you affect policy? Can you make a difference in policy? And the answer comes back. The answer is yes but only if and that if is if you have all the right relationships in place. So let me show you another what I think is a really cool project which is about pollution in the state of Gujarat. So and the title of it is truth telling by third party auditors and the response of polluting firms. The basic idea here is that the system there's regulatory, there are rules in Gujarat about pollution and there are inspectors. The inspectors and this is not so uncommon worldwide but the inspectors were being paid by the companies for which they were reporting on their pollution statistics and you can imagine what kind of impact that would have. I don't need to show you this everybody here knows where Gujarat is but but Gujarat so it's highly industrialized got a lot of the industrial manufacturing for this country. There's extremely high water and air pollution et cetera et cetera and it violates air quality standards. So the problem was that the Gujarat Pollution Control Board which regulates 20,000 of these industrial plants was regulating those plants A based on inspectors who were paid for by the company but also there was no negotiation with the company if you were above the limits this company would be shut down and if you're below the limits you continue to operate. So there was no real relationship and there was a bad incentive mechanism and this is just a picture of some inspectors I'm sure those are good inspectors. So this is a measure so before Epic proposed the experiment they wanted to propose this was the sort of official measure of pollution this is a bar graph showing the percentage of plants percentage of plants out of that 20,000 based on audits and the particulate matter that they were producing the Gujarat regulatory limit is about 150 milligrams per cubic meter and not surprisingly every single plant was producing at or below the limit and you could look at that and say great Gujarat's in great shape but of course when you look at pictures like this you ask the question can that really be true so the proposed experiment was actually this is an economist thinking about policy so first of all there's question of the data and second of all there was a question about incentives for the inspectors so what the proposed and this involved partnership between the Gujarat Pollution Control Board GCPB and the companies and what the proposed solution was had four parts first of all auditors would not be selected by the companies they'd be randomly assigned and a lot of this sounds obvious but this is an example of how you can make policy work first of all randomly assigned fixed payment from a central pool so they weren't gonna be paid by the companies anymore the auditors would be monitored meaning that once they made a measurement somebody else would come in a different auditor come in and make a measurement and if that first auditor was correct they'd get a bonus and if they were wrong they wouldn't so you know simple economics right basically incentivize the thing and sure enough this is what I showed you before this was this control and the new measurement looked like this so you know when you look at the sky of course it looks like this and this was just the first step this was truth telling this was basically saying I need good data so this is good data so the next step was to publish the data they published the data they had the government work with the companies and you know sure enough they made a dramatic change this is now audits after these are audits later so there's still companies that are polluting off but there's a much better percentage of companies that really are producing lower amounts of carbon in the end actually they were able to impact the total these are standard deviations from where they should have been but the total effect was about 28% so it's dramatic and very little was done they're not spending a lot more money all they're doing is they have a different system of incentivizing and now you have good data and you have good results so I'm showing this it's not a sophisticated data collection scheme it's not urban science in a deep way but it is the kind of science that works I guess that's the main point okay so I have a few more minutes so what I'd like to do is highlight a few of the other things that we're doing at the University of Chicago in the area of urban research and again asking that question what can research universities do in research but also in practice and many of these ideas since a lot of our work and we've been doing urban research like you have for many many years but our focus has really intensified recently and a lot of the things I'll show you are just starting I'll just give you a sense of the kinds of things we're thinking about but I also wanted to give you a sense of history so one of the more famous urban scientists actually in the world social scientists his name was Ernest Burgess at University of Chicago in 1921 he had a model for what cities should look like and the model was really based on Darwinian theory he basically said that cities were places where you had essentially competition amongst people and the competition was primarily for land but also for water you know the usual things we think of and purely based on theory not based on fact at the time he maybe slightly based on Chicago his assumption was that that people who could afford people with money would naturally desegregate out into these suburbs which as you know in the U.S. is the way that the way the cities look but not all cities he said that in the center would be the poorest people and then as you move further out uh... you would have the people who could afford to move and that was all based on a theory which was Darwin theory it was basically this competition theory now we know this is entirely wrong in fact even in Chicago it's not true in fact many of the wealthiest lived downtown if you've been in Manhattan I'm from New York you've been in Manhattan recently I mean it's you know the millionaires are complaining because the billionaires are buying them out you know it's just not the right model we know that and and Mumbai is is even more complicated I mean it's very mixed everywhere you go it's fractal I mean it's it's very complicated so what's interesting about it was this was a theory and this was the way urban science was done a hundred years ago or eighty years ago which I still find interesting nonetheless very theoretical what's changing now of course is that data is becoming something which which we have access to and and you know we've had you know we've had data for many years but not at the level we can finally start getting it we've had ethnographic data interviewing people and etc but what we're starting to get is real interesting data and this is just one example some of the things we have we just recently launched something called urban labs which is it which is um science which is basically data-based policy experiments essentially trying either proposing new policy taking policy that exists and doing essentially randomized controlled trials to see if they're working because most most strategies are not tested I mean government rolls out strategies that for a while in education that for a while in social welfare and poverty etc so we started to do is put together a set of labs in poverty health crime education and energy uh... and and started to to really think about quantifying and you know just like in the experiment I showed you in gudjarat start thinking about how we take some of the policy that exists and actually do good analysis and get good data this is an example of of one a set of data it's something called the crime lab and of course in an urban environment crime and education and nutrition etc all inextricably linked so you can't separate the two and this actually shows that here's a this is a bunch of it I won't go through all these these each of these is a different experiment and I won't I won't talk about them in detail but one for example of one summer chicago is a program now in chicago which takes kids in the summer and gives them jobs sounds sounds easy it's it is i mean it's not cheap but it's not very expensive but by doing this it turns out in these in the worst areas in these red areas which is where university chicago is we're just north of some of these uh... you actually get to reduce crime by forty three percent simply by doing something simple now you know social worker would have told you you do that and it'll help but this is quantifying really quantifying the results which is interesting uh... another one is called match which which is also very focused it takes african-american kids and in in grade seven through twelve the most vulnerable age in in in these in these fairly blighted neighborhoods and focuses on one thing uh... focuses on math believe it or not but it gives them it gives them extra tutoring in math typically these kids are three to four years behind the math brings them up to their age or their grade in math and it has a huge impact this is sort of a collateral impact very positive on violent crime arrests in those same neighborhoods so these kinds of control trials are very important kind of thing that we're doing another example which is much more technology and it gets back to some of the things i was mentioning before is something we call the array of things uh... this is worked on in collaboration with argon national labs which you mentioned so used to be director of that university of chicago something we call the urban center for computation and data uh... and the idea here is within within the summer this summer so we we only have a little data so far the summer were deploying a few hundred of these boxes and these boxes are going to measure a bunch of things they're going to measure thermal uh... wind rain etc which the usual things for weather but they're also going to range uh... measure carbon dioxide nitrogen nitrogen oxide pollutants uh... from automobiles they'll measure pedestrian traffic and they'll measure automobile traffic the question is what do you do with all that data and so i can give you some examples one i've already mentioned which is if you can measure these pollutants and then also do uh... health studies in the same region you can ask the question is there a connection between carbon dioxide and asthma for example because asthma as you know is growing dramatically in the states likely growing here too and it's likely due to a lot of the chemicals in the air this kind of thing you can say what we measure some of this today we do actually but not at the density we need so these things will be deployed on most street corners uh... issues are of course privacy because we're collecting gps data we're taking photos of every corner regularly throwing the photos out social scientists want to know where people are moving so what what's the human traffic look like so we're providing that kind of information so privacy issues are are a big deal but you know we had big article in the chicago tribune when we first announced this program big brother is watching you you know from george orwell's nineteen eighty four people are afraid that we're gonna somehow track them well it's not like it's not like your phone company doesn't know exactly where you are twenty four seven but nonetheless um... you know these are the kinds of things that we're thinking about here's another example so that's a good question we're gonna do it by neighborhood so we'll cover big areas of downtown chicago so let's call it you know fifty square blocks of downtown chicago ultimately we'd we'd like to cover a good portion by neighborhood the other thing they'll measure by the way is is things like energy consumption so you start to think about you know regional energy consumption which of course utilities have etc so so we plan to put these everywhere but we also wanted to put seven on our campus university chicago campus and our faculty got really upset because they felt we were invading their privacy so we finally convince them to do it but you know how that goes right sorry oh how much they cost that's actually a great question they're too expensive so far the the the chips are actually being built by intel so intel these are these are custom made chips so they're about six hundred dollars for the chip itself and the chip chip has like a electronic nose on it so it does some amount of the chemistry and then there'll be other kinds of sensors around it so it's about a thousand dollars a box which is too much we're also getting a deal with the the city has agreed to hang all these for free so there's also an in-kind contribution the city's making and you know you put somebody up on a pole these are um you know these are these are telephone poles right or whatever poles the utility poles uh... so it's not cheap to your point the goal of course is to reduce this by a factor of ten IBM is working on these two so eventually the price will come down but you need them to be if if they're a hundred dollars they're less than a stoplight and so you can imagine them being just part of a stoplight i mean to be part of that whole system but you it's exactly a question when you want you want a hundred of them in their thousand dollars that starts to become real money questions what to do with the data that's the fun part here's just another example before i finish because i will try to finish now this we we actually is another example of a cool program uh... this is funded uh... this is funded uh... just recently got funded actually by uh... got funded actually remember who this was uh... doesn't matter but this is a project funded called data science for the social good i'm not sure i love the title but but the idea was that every year now every summer we have a bunch of students come to chicago and proposed different types of of of experiments mostly digital based experiments around urban uh... urban uh... urban neighborhoods this one is an interesting one which is why i'm showing it uh... and uh... it's the city of memphis actually and like any city uh... the city of memphis has areas that are really blighted and abandoned some of them are banding really poor areas in the city and what they would like to do is ask the question where does uh... where does investment most pay off if you think about not just a municipal municipal financing but direct investment if we invested you know in in in this in this row of houses are the street how would we we optimally have a chance to have co-investment from uh... private money and so these students these graduate the college students uh... came up with an algorithm to actually calculate using some amount of economic theory calculate uh... what areas would make most sense uh... these are basically so green here is very low risk meaning if the city invested there it'd be easy uh... and in very high-risk red is where if you invest there there's no chance you're gonna regenerate uh... community uh... the city of course would like to ask where which corridor so you may not invest in red you probably don't invest in green because it's not very uh... it won't change very much so you start investing in light green or yellow but the point is this wasn't an economic analysis of neighborhoods it's an estimate but an idea that you could start thinking about uh... looking at the data in cities so you know data like what kinds of housing do you have uh... what kinds of storefronts do you have putting it all into one place this was an interesting idea basically how you take the stressed areas and try to improve them and they're actually trying so this is based on some students experiment the city of Memphis is going to try some of this in the next few uh... next few years of course working in urban on urban situations is really important and we do it because you know it's it's a it's a great challenge problem it's also the right thing to do but of course there's always the the the incentive which the technology i mean you asked how much those those units cost IBM is investing a whole lot in this i mean a lot of people are getting ready to think about not just smart cities and today we heard uh... we heard your prime minister announced this one trillion dollar program i guess it's a ten is a ten five or ten year program to invest a trillion dollars in smart cities a hundred smart cities so although i don't know for ten billion dollars you probably can't build a city so i guess they're really looking to co-invest but but uh... that announcement was impressive so maybe twenty five billion is too small a number maybe it's a trillion dollars but this is navi on which is one of the you know one of the consulting firms so it's not just that it's uh... it's a good idea people are going to invest in this there's money to be made uh... there's interesting not just solutions to social problems but they're huge social gains and one could imagine not just making money by building detectors and sensors but also thinking about using this data for marketing using the data for other types of things i think i'll finish there this is a city and good imagines that this is a city in gujarat it's uh... imagined as a smart city found this picture uh... recently there's actually a video to go with this about the smart city delirium do you guys did you know that city sorry so anyway it's just and i thought it was great way to end but you know the question of of smart cities aside because i think it's important to think about it in smart cities you know the the good news is that cities are very problem rich environments very interesting you know scientific problems but they won't be solved by science alone these are problems that you know that clearly involve scientists social scientists humanists lawyers uh... policy makers and none of these will get solved unless you have the right partnerships just not gonna happen and you know so the example in gujarat wouldn't have happened if michael greenstone didn't spend a lot of time building relationships with you know the the pollution control board industry getting help from actually talk to get the introductions he needed none of this would have happened it was really all that was as important probably more important than everything else so when we think as academics about solving these great problems we should try to solve them but we have to do it in a very embedded in clear way so with that i thank you all for listening i take any questions