Open cities are smarter & safer cities
An open data model to help DC reach its goal of zero traffic fatalities by 2024
More than 30,000 people in the US die each year in traffic related incidents; it’s a public health crisis. Vision Zero is a multi-national road safety initiative working with cities across the world to eliminate all traffic fatalities. The challenge for cities has been where to focus education, enforcement, and road engineering to make roads safer. In other words, where do you start?
We’ve partnered with our hometown of Washington D.C. to help the city reach its Vision Zero goal of zero traffic fatalities by the year 2024. Using open data on traffic incidents from the District Department of Transportation (DDOT) combined with anonymous sensor data, we’ve identified the highest risk areas to focus interventions — the results surprised us. We’re sharing the full story of this project today at Smart Cities Week.
About the analysis
In collaboration with DC, we began by examining the distribution of crashes in 3D, aggregated to census tracts and census blocks. At the block level, you can see that Union Station stands out as one particular hot spot of crash incidents.
To identify high risk areas for crashes in DC, we built a collision frequency model that compares several open datasets and sensor data. Open data on traffic safety has the potential to make our cities smarter and safer, but it must be paired with sufficient traffic volume counts for proper analysis — either from manual counting, automated methods, or sensor data.
We looked at the number of incidents normalized by the volume of vehicles and pedestrians and analyzed how incidents correlate to the density of businesses, schools, intersections, employment, census data, and driving speeds.
What we learned
Our initial hypothesis was confirmed — the more vehicles and pedestrians in a given area the higher the risk of injury. Lively urban streets with shops and restaurants attract more people and traffic, and as it seems, more accidents. Also as we expected, the more intersections on a stretch of road, the higher the number of crashes. However, the analysis did not find roads with higher observed speeds to have more crashes than those with slower moving traffic; roads with higher speed limits are more isolated and rarely lined with shops, restaurants, and pedestrians. Please note, we did not look at the severity of crashes in relation to speed limit, which has been confirmed before.
The output of the model uses all available data to identify high-risk areas for DDOT to better focus its Vision Zero efforts. The city can now take a more comprehensive, data-driven approach to re-engineer street design solutions, better support alternative modes of transport, reduce private vehicle usage, and revamp enforcement in the corridors that need it most.
Join us
We’re planning to test this collision model with more cities before we open source the code for all Vision Zero initiatives to benefit from this effort. Find me at Smart Cities Week, or connect with us through Mapbox Cities to learn more.
