After Accident Explorer's latest update, we started displaying accident clusters based on three-year intervals—but we were still curious about what clusters created using all the data, 2008-2014, would look like, and we thought you would be too. So we fired up a copy of Accident Explorer, removed the three-year restriction for clusters, eliminated all the clusters with fewer than 5 accidents, and started poking around.
Here are our picks for 15 of the worst places to drive in the United States.
This isn't a list of the largest clusters, although the number of accidents was a factor. We looked for a mix of rural and urban clusters, and for distribution throughout the United States, and ordered them roughly (though not exactly) by the number of accidents per cluster. If we'd gone just off of strict numbers, this would just be a list of densely-populated cities... and Florida:
(And remember that this view doesn't even include 3-4 accident clusters!)
To explorer fatal accident clusters yourself, visit Accident Explorer: https://accidents.apps.miosoft.com/explorer/
Be sure to read our other posts about Accident Explorer for information about how clusters are created, our data source, potential issues with the data, and more:
Post #1: Accident Explorer: Machine learning with traffic accident data
Post #2: Accident Explorer gets a turbo-boost
For Data Migration