Using big data to make a difference
In the mornings, Steven Koonin often dons a light blue shirt and khaki suit jacket, walks out of his apartment above Manhattan’s chic Washington Square park and heads for the subway. As he beelines down the sidewalk, the West Village buildings burp up black clouds of smoke as their boilers are fired on. At Sixth Avenue, an express bus screeches to the curb and blocks the pedestrian crosswalk. And as Koonin sits in the subway, he notices some of the signs are badly placed. “Can we fix this?” he wonders. He gets off at Brooklyn’s Jay Street-Metrotech station and rides an elevator to the 19th floor of a commanding building perched high above his native city. Then he gets to work.
Koonin is the director of New York University’s Center for Urban Science and Progress (CUSP), which is to say, he is the big data guru of New York City. He’s been given a lot of cash, millions of data points and a broad mandate by the city of New York: make it better. No big data project of this scale has been attempted before, and that’s because the tools never existed, until now. “There’s an enormous amount of data out there,” Koonin says with the vestiges of a Brooklyn accent. “If we can use the data to understand what’s going on in cities, we can improve them in a rational way.”
CUSP is both a research laboratory and a school. This year, it will have more than 60 students and 8 full-time faculty members. The students collaborate with the faculty and the city on big projects while they work toward either a Master of Science degree or an educational certificate. About a quarter of students this year will have social science degrees, another quarter each are engineers or scientists by training, and the rest will hail from fields as miscellaneous as film and fashion design. Their collective challenge is to turn numbers, spreadsheets, graphs and charts into a model that makes New York City work faster, cleaner, and more efficiently.
The program is already starting to make policy recommendations to the city, and as the institute attracts more talent, it will begin to play an important role in everything from easing Manhattan’s nasty rush hour traffic congestion, advising on prekindergarten school placement, cutting back on city pollution and helping businesses decide where best to open a franchise. “CUSP is able to work on those projects and take it to a deeper level of making more vetted recommendations,” says Nicholas O’Brien, the chief of staff in the Mayor’s Office of Data Analytics. “They bridge the gap between city data and creating actionable policy for city agencies.”
Koonin grew up in Bensonhurst, Brooklyn and in the late 1960s attended the prestigious Stuyvesant High School, where he and his friends once tried to use an old IBM computer (“it clunked along and had less power than your phone,” he says) to try and figure out the shortest time a subway rider could visit every single city stop on one fare. Koonin would go to the MTA headquarters to copy down timetables and input them into the computer.
Forty years later, Koonin has more data than he knows how to use. There are figures for household water consumption, purchases of goods, noise levels, taxi ridership, nutrition, traffic levels, restaurant inspections, parking violations and public park use; subway ridership, bus deployment, boiler lifespans, recycling rates, reservoir levels, street pedestrian counts; granular demographic breakdowns, household income, building permits, epidemic monitoring, toxin emissions, and on, and on and on. The challenge is making sense out of it, and that’s where CUSP comes in.
“The city has very little time to stand back and ask itself, ‘what are the patterns here?’” Koonin says. “That’s because they’re up to their asses in alligators, as you almost always are in government.”
Koonin would know. After receiving a Ph.D from MIT, he taught as a theoretical physics professor at Caltech before eventually working for BP and then the Obama administration. As Undersecretary of Energy for Science in the Obama administration, he was frustrated by the glacial progress on energy policy. To get things done, Koonin concluded, he needed a novel approach. “I came up with this notion of, ‘I’m going to go instrument a city as a scientist would,’” he says. In April 2012, he was announced director of the newly created CUSP program to make New York a living laboratory for urban improvement. Since then, Koonin has overseen a rapidly growing operation as it dances between 13 city agencies, the Metropolitan Transit Authority, the mayor’s office, and NYU, taking chunks of data and imagining actionable city policy.
CUSP’s temporary location (before it moves into the retrofitted Metropolitan Transit Authority headquarters) is an eclectic mix of high-tech and deep retro. The foyer, with firm orange chairs and dull wood paneling, looks like an Ikea designer recreated a 1970’s-era therapists’ office, but inside, two robots patrol the halls wielding touchscreens. A glass-enclosed conference room has 60 high-resolution monitors that on one Wednesday displayed the city’s taxi pick-up and drop-off data from the evening of May 1, and hundreds of teal and black taxi icons are scattered around a detailed digital map of Manhattan. In Koonin’s impressive corner office with magisterial vistas of downtown Brooklyn, he keeps a classic slate blackboard next to a keyboard. He can fluidly play “You Go To My Head,” the J. Fred Coots jazz standard, and “The Way You Look Tonight.”
“My dream is to be a lounge pianist,” Koonin the data-meister says drolly.
Like a doctor holding a prodigious stethoscope to New York City’s skyscrapers, Koonin needs to give the city a thorough physical before he can write a prescription. “The city has a pulse, it has a rhythm. It happens every day. There’s a characteristic pattern in the rise of economic activity, energy use, water use, taxi rides, et cetera,” Koonin says. “Can we measure the physiology of the city in its various dimensions? And define what normal is? What’s normal for a weekday, what’s normal for a weekend?”
“Then you can start to look for abnormalities,” he continues. “If subway ridership was low, was that correlated with the weather? When subway ridership is low, is taxi ridership high? You get a sense of what’s connected to what in the city. Can we look for anomalies, precursors of things? Epidemics, economic slowdown. So measuring the pulse of the city is one of the big things we’re after.”
CUSP is creating a system to measure microbiological samples from the city’s sewage system, using genomic technology to learn more about people’s nutrition and disease based on their waste. Do certain neighborhoods need better nutritional or hygienic practices? Another project involves a camera fixed to the roof of CUSP headquarters that can see anonymized data of when people’s lights turn on and off and monitor energy usage. When do people go to sleep? How regular are people’s sleeping hours? The institute is also working out a way to help the city’s Parks Department measure how many people use city parks, and what they do in them. (Hint: it could involve lots of infrared video.) The city could then much more intelligently design its public spaces.
“This is opening the door to the possibility that we would very accurately and very comprehensively understand how people would use our public spaces,” says Jacqueline Lu, director of analytics at the Parks Department.
The city’s 8.3 million-strong crowds, packed together on the subway like brightly colored gumballs or streaming through the streets like grains of sand blown by the wind, will be the ultimate beneficiaries of Koonin’s work. On his morning commute, he notes how the city has changed since he was a kid coming up in the public schools. “Everyday it’s really interesting to look at the crowds and see how they interact with one another,” he says. “The city works better. The trains are pretty much on time. So it’s pretty good.”