Full Interview: Tracking Transportation in New York City During the Pandemic with Juan Francisco Saldarriaga
By: Hannah Lin (CC '23)
Can you talk a bit about your background and where you work?
I have degrees in architecture and urban planning from Columbia, and I also have a B.A. in philosophy. I am currently the Senior Data and Design Researcher at the Brown Institute. We are a bicoastal organization, both at the journalism school at Columbia and the engineering school at Stanford. We were funded by a gift by Helen Gurley Brown, who was the longtime editor of Cosmopolitan magazine. We try to combine technology and tools with reporting and storytelling. Every year, we give around 1 million dollars in grants to different teams that are doing interesting things combining tools, technology, storytelling, and journalism.
You research at the intersection of so many different fields. Those are, and I’m quoting directly from your website, data, geographic information systems (GIS), visualization, journalism, architecture, urbanism, and the humanities. How did you get into this work?
I first got into GIS, geographic information systems, and mapping in my urban planning masters. When I graduated from that, I started working in the Spatial Information Design Lab, which was a lab at the architecture school at the time. There, we were doing research on data visualization, working specifically with datasets around cities and transportation: urban data. Then, the Spatial Information Design Lab became what’s now the Center for Spatial Research after we received a big grant. I was one of the main researchers there.
Part of what made the Spatial Information Design Lab special was that we often collaborated with people from other disciplines. That included the Brown Institute for Media Innovation, where I work now, because a lot of the ways in which we were trying to present information and data overlapped with the work that journalists do in telling stories.
We also collaborated a lot with people in the humanities because a lot of the datasets we work with around cities, population, and migration are also humanities datasets. There are also travel diaries. There are also letters between people who are corresponding from one city to another. There are also datasets about forced migration. For us, it was crucial to not only get different perspectives in terms of how we were telling the stories but also in terms of the actual topics, so we needed to collaborate with experts in the field.
Can you describe the research you’ve been doing related to the pandemic?
We’ve been doing two things. One is a more general Brown Institute initiative. In early March we decided to do a rapid call for proposals around COVID and journalism or research. We put out the request for proposals and set a deadline of ten days for submission and evaluation. We had a very fast turnaround; we wanted to give our response back in three or four days.
These were for microgrants, so we were giving out around 5000 dollars. We had quite a small budget, around 25,000 dollars, but we had an enormous response. We had more than 300 proposals combining journalism, storytelling, and data visualization around COVID—proposals from local newsrooms, from developers, from NGOs, from community organizations.
There was a very interesting proposal that wanted to study domestic violence during COVID, because the hypothesis was that we would see much more domestic violence. There was another one that wanted to include machine learning in how they process questions from the audience for a radio station so that the responses could be more efficient and automated, freeing up time for journalists. There was another very interesting one from a radio station in Alaska, where they wanted to just get equipment so that they could transmit their COVID-related information better to a broader audience—it was the only radio station that covered a big swath of land, and without that, the community didn’t have any information. Our teams got together for a week and read all the proposals. We were happy to give out the grants, which have already started producing good results.
On a more personal level, I am currently doing a research project around transportation and COVID in New York City in partnership with David King, who is a researcher at Arizona State University. We are looking at how the choice of transportation mode for New Yorkers has shifted throughout the crisis. Initially, we wanted to test whether there had been a similar decrease in usage of all transportation modes, especially subway and Citi Bike. Our hypothesis was that people actually switched to Citi Bike for a little while before everything shut down, so we wanted to compare the decrease in both transportation modes.
We’re still working on it, so we haven’t published it or written it, but what we found was that both transportation modes decreased simultaneously in a very parallel manner, in a similar magnitude of decrease. They didn’t decrease from one day to the next; it took them over 8-10 days to really shut down, but for both of them, if you see the curve, they really mirrored each other.
What’s been interesting has been the reopening. We have actually seen that the number of riders taking the subway has really not picked up that much, whereas the number of people riding Citi Bikes and moving around on bicycles—the city has bicycle counters on the bridges—has picked up. Right now, we’re in the middle of testing whether they have picked up to normal levels, or whether they’re still below normal levels. That’s one step we still have to figure out.
We also still need to figure out what kinds of trips these are, because these could be trips of people going to work that have replaced subway trips, or these could be different kinds of trips: a person locked at home who decides to go for a ride after lunch, a person who decides to bike to the supermarket. We want to understand what kinds of trips are the ones that have picked up, if it’s a substantially different kind of trip or if it’s a commuting trip and people are actually replacing subway trips with bike rides.
Finally, the other thing we want to understand is the spatial distribution of these changes. There have been a few good analyses looking at the decrease in subway usage and how the decrease, even though it was enormous (more than 50%), has not been uniform across the city. Some neighborhoods have seen decreases around 90%, while others have seen decreases around 40%. That’s pretty much tied to demographics, especially income and people who are deemed essential who are, most of the time, in lower-income neighborhoods that are still relying on public transportation. We want to understand what the spatial distribution of that decrease is and what the spatial distribution of this new rise in Citi Bike usage is. Citi Bikes, of course, don’t cover the whole city, so how does that play out compared to the subway?
How exactly do you go about getting the data, especially if you’re trying to distinguish the commute type, between going to work or the supermarket?
All of these datasets are publicly available. For subway data, you can get it by station and you can get the number of entries and exits per station, but you can’t really tell the type of trip. You can tell that a person entered in one station and that another person (or the same person, who knows) exited in another station.
Citi Bike does make their detailed data available, so that means you can see each individual trip. You cannot track users across the system, but you can see that this trip started in this station and ended in that station. In the case of Citi Bikes, you can tell the duration of the trip, the time of day, and the origin and destination neighborhoods, and you can infer whether those are commuting trips or not by the time of day in which they happen and their origin and destination—whether it’s a residential neighborhood as an origin or an office neighborhood as a destination.
For bicycle counters, you can’t tell what type of trip it is, but by the location of the counters, you can infer some things. Let’s say you see a lot of Citi Bike activity in Brooklyn but you don’t see that much activity on the bridges, where the counters are. You could infer that there are a lot of people biking, but not necessarily biking to work. All three datasets are public and available and are updated on a fairly regular basis, but they’re not exactly the same in terms of the amount of detail they give.
Will this transportation research project be going on for the foreseeable future?
I want to get it out soon. This is the thing with academic projects—they always take way, way longer than you want. Ideally, it will go out to an academic journal and, at the same time, to a news outlet, just because I am in the journalism school. We still have to write it up and see if our conclusions are interesting and novel and worth publishing. Hopefully, they will be, but you never know.
What projects are you looking into for the coming future?
There are a few research projects that I have in the back of my head. There is one urban planning project looking at segregation in cities, especially in Colombia, the country where I’m from. Cities there are divided by income levels. There’s a map that actually says, “this neighborhood is middle income, this neighborhood is low income,” and depending on that designation on the map, you pay different rates for your utilities and different taxes for your property. It’s a sort of government-sanctioned economic segregation. I’ve been collecting Twitter data on how people talk about the segregation for around two or three years, and I want to do an analysis where I test how much this policy has influenced the way people talk about segregation.
There are other things we do at the Brown Institute. As I said, we give out a million dollars in grants. In this year’s cohort, a lot of grants have to do with COVID or with the protests around the killing of George Floyd. One of my main jobs is to accompany these grants and help them throughout the year. We have a grant that is looking back at the 1994 crime bill and the Central Park Five case and looking at how the media was talking about crime in that moment to understand the role it played in racializing crime. That’s one of our biggest grants, with a team from Columbia and Stanford, so a bicoastal grant, which we really like.
There are a couple more. There is a very interesting one going on right now, in which we are helping a researcher submit FOIA requests to local or state agencies around COVID, so requesting those agencies to give us everything they have that mentions COVID. My boss says that we’re writing the first draft of history by seeing how city and state administrators are dealing with COVID. The researchers in the grant have uncovered very interesting ways in which different cities have approached the pandemic.
For example, there was an article in The New York Times about a month and half ago about how New Orleans was preparing for COVID, especially around Mardi Gras, and the way they were able to tell that was from the emails that we got. There was another series of articles around meatpacking plants in Kansas and how the state officials knew that the plants were big hotspots, but they kept that from the public. That’s a very interesting project around COVID that’s happening right now on which I am working as an advisor, helping the grant group.
What is your perspective on how the pandemic is shaping the future of journalism and data?
From what I understand, journalism is in a very important moment. Before the pandemic, it was in a difficult financial position. However, with the pandemic, with all the other things going on, the value of journalism is becoming more clear and people are starting to realize that supporting good journalism is important not only for themselves but also for a healthy democracy. I’m hoping that that really serves to maintain the good journalism that is happening and to encourage more of it.
I think journalists are becoming very good at working with data and at questioning data. Data is an incredibly powerful resource, but it’s also something that needs to be questioned and well understood. You shouldn’t put all of your trust in a dataset. Journalists are becoming very good at doing that, so I am hopeful. I see a lot of experimentation happening in news outlets, a lot of experimentation in visualization, which is great. I would say they are leading the way in experimenting around visualization, and that’s something that everyone is benefitting from. I’m very happy to be in this area at this moment, because I think it’s very exciting.
For students who want to head toward a similar career path that you’re on, do you have any advice or life lessons that you want to impart?
I often wish I had studied computer science and had more of a background in the more technical aspects of my work. That would make it much easier. At the same time, I value my training in urban planning, architecture, and philosophy for giving me the conceptual foundations and a critical eye, both in terms of design and also conceptually in how to understand problems. I see a lot of analysis done just because someone knows how to use a tool, but they haven’t really thought about what it means or what kind of analysis is the appropriate one.
I would encourage trying to get both, if possible. The technical aspect is something that can be learned on your own, which I did, so that’s possible and that’s great, but a more rigorous technical training would be better. A rigorous training on a subject matter, like urban planning, architecture, or the humanities, is also very valuable. So, hopefully, get both!