B (3:39)

I have a lot of questions for you in that regard and a project that we're working on at the Bloomberg center for Cities. But before I get there, one of the challenges that I've seen most just in listening to your explanation, is that there's no one domain, no one agency that has control of the data. And we've wrestled a lot, even within a city, about how folks in the environmental department, I'll misname the departments on purpose, can see data from public health sources in order to concentrate on interventions or preventions. So how do you think about the agency issues and data access issues? Then we'll go to some specifics.

C (4:26)

Well, I mean, I think you're right that we have often fragmented data infrastructures in public sectors throughout the world and US Cities as well as many others. And that's the, I think, the result of, like, incidental architectures for data. Things grew up around, you know, areas of, you know, kind of localized responsibility, agency localized responsibility. But we're not designed for interoperability. We're not designed for transparency and accountability, and we're not designed for modeling that is in need of inputs from many different sources. So to your point, that lack of intentionality results in not really having a integrated data architecture. The rise of the data stewardship movement, I think, is one response to that. So, you know, we work a lot with Dr. Stefan Verholst at NYU in the governance Lab around how to intentionally correct that, which means creating jobs that have this in mind and that crossover also to the private sector. So increasingly, you know, the data that's most needed being produced at largest scale is not being produced by the public sector. It's being produced by a range of private actors. And there the situation is even harder because the incentives often don't align. The legal agreements are more fraught, they're more complicated, privacy controls, et cetera. And, you know, again, you have the lack of intentionality. So we, we need designed for that. Chief data officers, I guess, are one solution to this. But you need to go far beyond this kind of process and really focus on making sure that people practice data integration as well so that you can understand what changes when you get it right.

B (6:20)

Let me discuss with you just for a second, a project we're working on. You're going to actually be involved a little bit with this project and then come back to the conversation we're having now. One interesting thing is what is a crisis? I mean, it seems to me like continuous exposure to problems which we know are going to create bad results and they could flare up. And you could call the flare up a crisis, or you could call every day a crisis because you're not dealing with it. And come back to that in a second. We have this Robert Wood Johnson foundation initiative called the Community Data Health Initiative. And in that we're looking at the local level, so city level, and trying to identify what are the pollutants that cause public health adverse effects and about which the city could do something. So how do you think about PM2.5 or carbon and chronic childhood asthma, or lead and brain development? How do you use the data to identify and intervene? So let me just use that for a second. Talk a little bit in terms of the sophisticated work you do, how you could use data to understand threats such as this and responses to those threats.

C (7:38)

It's sort of an old adage actually for us in disaster response that the places most at risk to the flare up or the crisis, as you put it, are the places that were most at risk the day before. And that's actually a complex interconnected problem. So, you know, when an urban wildfire hits a place like Los Angeles, you know, the disproportionate effects fall on people that actually have health risks that emerge from the hyperlocalized environment that precedes the outbreak of the fire. The fire releases a significant amount of new particulate, new pollution, and that does have its own hyper localized environment based upon airflow, based upon the built environment, et cetera. But it lands into a landscape of unevenly distributed health risks. So that's something which we've had to understand and I think increasingly granular and increasingly sort of like biologically complex ways. So I think it wasn't that long ago actually when we were mainly talking about things like wildfire smoke risk in terms of respiratory illness. But that's actually only a small fraction of what the kinds of particulate releases that we see all the time are actually affecting. This is affecting the blood brain barrier, this is affecting dementia risk and Alzheimer's. This is affecting cardiovascular illness. This is affecting long term risk for diseases including things like diabetes and the endocrine system. So we have to be able to understand that at a hyperlocal level in the sort of physical system of the body, but also across neighborhoods, because that uneven exposure is based upon, you know, the uneven exposure of where sites for pollutants are located. It's based on uneven access to quality respiration. Right. Air quality, access to air conditioning is a significant issue. And so when you put that into some kind of model for looking at what's likely to happen in a crisis, you actually have to feed that environment that you're talking about into that model or you're not actually going to even understand the crisis when it happens.

B (9:53)

So let's do this. Let's say that we're all motivated by your last answer and we're going to go tell three or four cities how to structure this to produce the best possible response. So you've got data, you've got multiple agencies, you have this preventative issue, you have these kind of latent. Maybe latent's the wrong word. Conditions. So give us an example. May be responding to my Robert Wood Johnson foundation description. But how would you set up that structure? I mean, your comment about the NYU work is interesting. How do you put that together? What should we tell mayors?

C (10:33)

Well, I think you need a combination of platforms, data and people. And mayors need to be able to organize all three of these things, and they need support from probably beyond just the public itself. So, you know, the platform issue I think is interesting because we now live in a time where data platforms are actually much better than they've ever been. And this is like, you know, the cloud data revolution that is pioneered by companies like Snowflake and Databricks, Google BigQuery, Google Earth Engine, et cetera, et cetera. There's, there's more capacity now for being able to, to really efficiently host and integrate large amounts of data than ever before. Some of that is public APIs, some of that's private, you know, et cetera. But in terms of the ability to use that, I think that's actually something which is a new kind of opportunity. The data is something which, you know, requires the kind of clear understanding of what problem you're trying to solve. To your Robert Wood Johnson point, you know, what are the sources where we're actually going to understand microclimates for particulate, for pollutants, for temperature. Micro variants of temperature make a big difference in terms of things like heat exposure and other risk factors. And that's coming through satellite imagery, that's coming through sensors. There's networks like Purple Air which are providing respiratory information. There's was just at the meetings in D.C. for the new Firesat project, which is initiated through Google Research, but is running now through the Environmental Defense Fund and others. This is going to be a global constellation of fire detection satellites and also measuring things like local heat exposure, things like that, smoke plume modeling and PM2.5 concentrations, including the height. So that has to be something that people are aware of, that they're able to handle. And that means that you actually then as the mayor's office need to recruit data officers and scientists as part of the core function of what the public sector means right now. I think that's a challenge actually for a lot of places. Like they may not all understand, you know, the art of the possible. The, the ability to actually recruit people into those jobs is really important. And that's also, you know, an area where they're in competition with other actors. So how we kind of think through the economics of recruiting data staffs that are able to do that is also a central public policy issue, you know, and, and then you can sort of put ethics and privacy and other sort of concerns around that so that you're maintaining public trust because you can get the data right and do it in the wrong way and lose the trust of the public and then you're not meeting your political sort of priorities and considerations and then you might actually lose support for other kind of analytic efforts. So making sure that all of that is functioning is the new job of the mayor, I suppose, but actually pretty exciting, I think.

B (13:32)

I think, yeah, if a mayor did this, they would need a platform provider, probably a nonprofit or university or academic center mayor could convene. But there needs to be a platform provider, I think, in some sophisticated sense.

C (13:49)

Yeah, I mean, that's one of the things we've been kind of putting together through the Crisis Ready project. We did a project called climateverse which is being rolled out now in India. We're going to be testing this out in Mexico City in a upcoming workshop in November. The idea of climate verse was that the data required for cities and states to be able to do effective climate modeling. Climate analysis related to the kinds of micro environments that you're talking about is out there, but is often not easily accessible in the sense of understanding the data limitations, the missing data, the, you know, just sort of technical access considerations. And then we use some LLM technology just to be able to help people to have queries in conversational form. You know, so often you'll have cities where the right people are not really trained data scientists. They have questions, they don't necessarily know how to frame something in the way a data scientist would. AI can help with some of that and can help you then Understand what data is there to be able to solve that problem. And that's where good projects come from. Expanding that out is really one of the things that I think we can do a lot of. And that's really kind of geography independent. I find I've not yet met a place that doesn't need something like this. So a lot of work to do.

B (15:22)

Very helpful. I just have a couple more questions, but let me just stay with this for a second. I think you could look at this. I just keep using the Robert Wood Johnson as an example. I think you could do this in any number of areas that lead to crisis, but you could chase the health results and identify a hotspot for childhood asthma, chronic childhood asthma, and. Or you could chase the point sources of the pollution, which, you know, are connected to certain conditions, and then look to see if there are conditions. What's the best way, though, to put together the health and the environmental data, whether you're thinking nationally or locally? Because that seems like, I mean, and for every good idea, there's 10 lawyers that tell you how you can't share the data.

C (16:16)

Yeah, I mean, I'm not a lawyer, so I actually don't know entirely how to solve the legal problems with this, except to note that you're right. It is something that requires really clear, consistent responses from a legal standpoint. I mean, I think to answer your question from my point of view is to make sure we're clear on what question we're asking them. What is the mode of redress that's possible? So if you're asking a policy question about mitigation, that's one constellation of health and environmental data that's about being able to regulate polluters, that's being able to understand what's in emissions, that's a whole bunch of industrial regulations and transport regulations. That's. That's a different question than saying, you know, if people are going to be showing up at, you know, clinics and hospitals with certain kinds of conditions, and we're going to need to think about medical resources for those people in a way that's mindful of the risk landscape that they're embedded in, what are we going to expect? And that I think actually just requires. It requires a similar, but actually different kind of, you know, sort of data constellation to be able to answer those questions. And the agreements are going to be different and the level of kind of personal detail is going to be different. And we need to be able to have frameworks for doing that kind of analysis, for constructing those data pipelines. That are mindful of context and that are portable depending on sort of which question is being addressed and what we want to do as a result of that question. I mean we usually ask that in the crisis ready context when we're engaging with cities and other actors as to let's focus on not just what you want to know, but what you want to do with that knowledge. And that is usually a pretty good guide to how to be able to construct the frameworks that you're talking about.

B (18:15)

These are such great answers. I could utilize about a thousand hours of your consulting time, but maybe you don't have that much time. Let's do this. Let's close with a preview. So you're going to during Climate Week, lead a discussion with some number of cities in the Salad Institute, Bloomberg center for Cities on Heat and Heat's an interesting subject, right? Because we heard from the Chan School about heat and low birth weight babies. We've heard about heat and all sorts of other pretty serious developmental problems. We've heard about heat and how it reduces test scores in classrooms. And this would seem to be a perfect use case of what you're all about, right? Bringing the data to bear on an issue which is a crisis and about which I would suggest if people paid more attention, they could actually help make it a little bit better. So give us a two minute preview of what you're going to do that day to motivate our audience to pay more attention to these issues.

C (19:19)

Well, first of all, thank you for helping to draw attention to it. So what we're going to do on the 18th is modeled on work that my colleague Rachel Balsari from the medical school and public health school and Robert Mead and Tess Whiskel, you know, heat physiologists and medical doctors really put together for an event that we ran for Harvard faculty back in the winter, looking at what Dr. Balsari describes as a kind of logarithmic scale of crisis for heat, where we need to understand the built environment, the way that that's situated within heat landscapes, the way that, you know those policy issues leading towards why you have certain kinds of exposure dynamics and vulnerabilities is there, but then really immerse people in the cities. In this case, it's people that are leading public health and crisis response efforts in a number of different cities like Boston and Jacksonville and Tucson, to play the roles of different actors that are required to coordinate to be able to effectively respond to this. And this is something which is just starting to emerge, I think, through the leadership of cities like Phoenix, Arizona and Miami, Florida, where, you know, we need people that are able to not only understand the unhoused population, people that are much more exposed to extreme temperatures just because of their physical situation, but also, you know, the data sources for being able to track those exposures, the people that are going to be dealing at the clinical level or at the hospital level with heat shock and with the kinds of emergency medical situations and really test out effective dynamics for coordination during a crisis event where you'll get sort of like an extreme spike up to, you know, really dangerous levels for a certain period of time. How do we understand that duration, nighttime exposure, et cetera. And those coordination efforts are, I think, eye opening for cities to be able to really the joint effort required to mount an effective response. So that's what people are going to be put in the middle of and then reflect upon so that we can then port that into their planning efforts in their cities. And I think it should be a pretty good time. We have a lot of video that we've kind of prepared for this to try to like put people in the simulated mindset of this is happening right now. So that's, that's what we're doing on the 18th. You know, in the afternoon on the 18th, we're going to be also doing a session on data preparedness and urban disasters and AI elements are going to be part of that. So we're going to do a series of lightning talks, for instance. One of them is a group from Texas A and M that's using drone technology that is loaded with computer vision models to do real time damage detection on the drone for flooded areas or places that have been burned, things like that. So that's like an interesting example. We have Overture Maps foundation joining us, which is large scale integrated open data. A lot of that inputs into AI models for doing large scale damage detection. We use it for like Hurricane Otis in Acapulco where we were able to do damage assessments within 48 hours at scale on the entire city. That turned out to be quite accurate actually. So thinking about resource planning, think about integration of that into city response and planning workflows, things like that. So I don't know, there's like a computer vision might be the sort of main way I would think about that.

B (23:03)

That is very exciting. I w. When I was deputy mayor of New York, the Office of Emergency Management was in my kind of organizational chart, this is about a decade ago. And they spent a lot of time figuring out how to move physical assets to places of crisis. Not so much time trying to figure out how to move data in order to understand the crisis. Right? So your addition of this idea, a tabletop for data, actually exercise, is quite exciting. Andrew, I can just think of so many applications. Let me just thank you for your time, your work, direct relief, crisis ready, and helping the cities around the world do a better job. This is Steve Goldsmith, professor of Urban Policy at the Bloomberg center for Cities at Harvard University, thanking Andrew Schroeder for his time and his efforts. Thank you very much.

C (23:54)

Thank you. Really appreciate it.

A (23:56)

If you like this podcast, please Visit us at datasmartcities.org and find us on itunes, Spotify, or wherever you get your podcasts. This podcast was hosted by Stephen Goldsmith and produced by me, Betsy Gardner. Thanks for listening.