Cities working to reduce traffic deaths are missing potentially valuable troves of data that our cars generate every day.
A group of New York University graduate students found that the computerized data from most cars on New York City streets — tracking information like hard-braking and sharp speed increases — could offer a better understanding of driver behavior and add another layer of insight into identifying dangerous streets ripe for improvements.
“Having granular data like that you can start to look at driving behavior in much more in-depth ways than we do now. Really the only ways we can measure speeds is with a radar gun or hoses that cross the road,” said Ariel Kaputkin, one of the graduate students who authored the report “Using Real Driving Data For Transportation Policy.”
“What’s really powerful about this data is it’s actually attached to an individual car, so you can track cars around turns, or down several blocks,” kaputkin said.
The students teamed up with the smart-driving app Dash to analyze data the company collected from the computer systems in 281 vehicles over a six-month period. While the scope of the proof-of-concept report is too small to produce conclusive findings, Kaputkin and his colleagues were able to map out speeding based on time of day, weather, posted speed limits and location. Speeding was most often identified on thoroughfares like Henry Hudson Parkway along Manhattan’s Upper West Side and Brooklyn Queens Expressway, among others.
The report also questions the efficacy of speed humps, finding that drivers tended to quickly speed up immediately after passing over humps. Compared to nearby streets without them, mean, median and maximum speeds were similar, according to the report. And parts of lower Manhattan, midtown, and Brooklyn Heights were highlighted for hard braking, which could be used to identify “near-miss” collisions.
“This is a very interesting way to get a sense of how drivers are actually behaving rather than how the city wants them to behave,” said Sarah Kaufman, assistant director for technology programming at the project’s adviser NYU Rudin Center for Transportation. “Using this kind of granular data lets us see how drivers are behaving, say, before a crash, rather than after a crash has occurred to gather data on what might be an intersection that could use a redesign ... So if we can track near misses, rather than [crashes] we can actually change the street design before a catastrophe occurs.”
Dash, a New York-based company, offers real-time feedback on your driving skills, trip tracking, and access to vehicle diagnostics by tapping into computerized data within modern cars. Dash is a practical tool for car owners, but Jamyn Edis, the company’s CEO and founder who also works as an adjunct professor at NYU Stern, said the company is looking to responsibly and productively monetize its data, say, by helping the public sector improve safety.
Dash worked on a pilot with the city’s Department of Transportation in 2015 that incentivized good driving behavior by offering discounts on insurance.
“There are about 100 million lines of code in the average car now that gets sold — which is pretty staggering. All the data is being gathered and processed from censors around the vehicle and that’s how it keeps it running,” Edis said. “But it’s ephemeral. It might show up on a couple of gauges on the dashboard but the consumer has no way to access it.”
Tracking all the private vehicles in the city raises obvious privacy concerns. But New York City oversees a fleet of 30,000 municipal vehicles and there are another 130,000 vehicles licensed by the city’s Taxi and Limousine Commission. In the city’s Vision Zero Year Three report, published early last year, the de Blasio administration outlined an initiative to track the city’s municipal fleet with similar technology to “gain an unprecedented amount of information about driver behavior.”
It’s unclear what progress the city has made on the initiative. A DOT spokesman said the program is still in development, but that the city plans to use the data for safety, planning and research purposes.
“I do think that maybe the general public would not be on board with this, but there could be incentives offered, like discounts on insurance rates,” said Kaufman. “There is a discussion of personal data, but the city tracking the data from its own cars is not an invasion of privacy.”