Data scientists target serious crashes in four U.S. cities

By Chris McCahill

Microsoft has teamed with DataKind, a nonprofit organization that connects communities with skilled data scientists, to reduce traffic-related deaths and serious injuries in four U.S. cities. The effort, which first launched in New York City last August, will expand to San Jose, Seattle, and New Orleans.

The DataKind Vision Zero project will harness public and private data to understand and prevent serious crashes in each city. Beginning in New York, the team will evaluate different Vision Zero strategies to determine the most effective approaches and inform future investments. They will also leverage that knowledge to inform decision making in the other cities.

In San Jose, the team will use learning algorithms to understand patterns of crashes in order to prioritize traffic calming investments. In Seattle they will analyze how driver behavior and street design contribute to bicycle and pedestrian crashes, so that the city can properly allocate funds from its Move Seattle levy, passed last November. In New Orleans, they will evaluate whether 911 call data can serve as a real-time proxy for bike and pedestrian crash data.

Although crash-related factors have been studied widely in research, public agencies rarely have the resources to replicate those studies or the volition to act on the findings. With buy-in from local agencies, these projects could help push data-driven decision making to the forefront.

The projects will be led by Michael Dowd, DataKind’s Science Fellow, with support from Microsoft’s Technology & Civic Engagement team.

Chris McCahill is a Senior Associate at SSTI.