But the big question about those efforts is just how effective they really are. Now, new data is emerging that gives policymakers a better picture of where Vision Zero is working. And more data tools are on the way that could help address dangerous conditions before traffic deaths or injuries occur.
One recent analysis of Vision Zero in New York City found that the improvements did seem to be working. Alex Armlovich at the Manhattan Institute, a conservative think tank, concluded that pedestrian and cyclist deaths at intersections with safety treatments decreased by 34 percent from 2009 to 2016. That far exceeded the 3 percent decrease during the same time span for intersections that did not get new safety features.
Overall, pedestrian deaths in New York City last year increased slightly, from 126 to 128. But the number of pedestrian deaths at intersections with Vision Zero improvements actually decreased during that time; the increase came from places that did not get the new features.
“Vision Zero’s progress is visible even at a coarse neighborhood level,” Armlovich writes. “The concentration of treatments in a given community district has a strong relationship with the decline in pedestrian crashes.”
Lower-income neighborhoods remain the most dangerous for pedestrians and cyclists, who are 9 percent more likely to be injured or killed in traffic accidents in the 10 poorest New York neighborhoods than in residential neighborhoods as a whole, Armlovich notes. The city hasn’t deployed Vision Zero improvements to lower-income neighborhoods, in part because of local political opposition, he says. “Change is slow and difficult for one main reason: any change, even if a majority of constituents favor it or are indifferent to it, upsets the status quo, and the status qo often benefits politically active residents,” Amrlovich writes.
“Though Vision Zero NIMBYism occurs in rich, middle-class and poorer neighborhoods, delays to safety fixes in poorer neighborhoods mean that they remain much worse off than wealthier neighborhoods, as they start off at such a disadvantage,” he adds.
In one respect, though, Armlovich’s study demonstrates the limitations policymakers face when they rely on data to make traffic safety decisions: In order to get a reliable sample of data to draw from, analysts have to either look at several years’ worth of information or, as Armlovich did, use broad categories of data (such as treated vs. untreated intersections), which makes it harder to quantify the effects of specific changes. It is much harder to quantify the effect of changes at a specific time and place. And, of course, researchers have to wait for someone to get hurt or killed to get the data they need.
But that, too, might be changing.
In another recent study, researchers from New York University compared police reports on collisions to driver behavior data collected by Zendrive, a private company that uses mobile phone apps to gather information on dangerous driver actions. NYU’s researchers found that the risky behavior corresponded to actual collisions 71 percent of the time.
The data from Zendrive, however, was a lot quicker to collect than the information from the police itself. NYU drew on just six months of Zendrive’s data for its analysis, compared with nearly five years’ worth of police reports. What that suggests is that analysts can get a more up-to-date – but still accurate –picture of traffic conditions using data collected by mobile phone apps.
(How does Zendrive get its data, exactly? It was initially just an opt-in app used by commercial fleet owners, which helped them manage their vehicles, dispatch drivers and get discounts on insurance. But now it includes everyday users, too. Its movement-tracking capabilities are included, for example, in an app that will immediately notify a user’s family if the user is in a crash. The one caveat is that Zendrive’s phone app is designed for normal passenger vehicles – not trucks or buses – because it is not designed to account for how those larger vehicles move.)
Noah Budnick, Zendrive’s director of public policy and government affairs, says the type of data the company is collecting will enable public officials to make quicker decisions. He notes that in San Francisco, Zendrive’s data quickly showed an ambitious project to remake Mission Street is already working. The Municipal Transportation Agency redeveloped the corridor to make its bus service there more reliable, to reduce traffic and to make the street safer. The new configuration included two dedicated bus lanes (each painted red) and new turn lanes.
Based on data gathered during 100,000 trips over the first 10 months of last year, Zendrive found a 16 percent reduction in “risky” behaviors by drivers after the Mission Street redevelopment was completed. That included a 36 percent reduction in excessive speed, a 30 percent drop in aggressive acceleration and a 21 percent decrease in hard braking. Risky phone use, though, only fell by 8 percent.
“In essence, the SFMTA took a chaotic commercial corridor with poor bus service and transformed it into a bus priority route to improve the experience of the majority of people who travel and shop there – walkers and transit riders,” Zendrive reported.
Budnick acknowledges that real-time data won’t be the only information policymakers need, but it at least gives them instant feedback. “When I hear from traffic engineers, they say there’s no such thing as too much data,” he says.
For now, Zendrive only shares its data with local officials for special projects, like the NYU analysis. Local governments still vary widely in their ability to process and use this type of data, Budnick says, and both the company and outside researchers are still figuring out different ways it can be used.
But the NYU researchers say the data could improve Vision Zero and other safety campaigns. It could be used not just to decide what improvements to build, but also what education and enforcement campaigns a city ought to undertake to save the most lives.
“The more detailed analysis, the more precise, actionable steps communities can take to achieve Vision Zero,” they wrote. “This data is invaluable to all Vision Zero stakeholders.”
