DC Releases Huge Trove of Crash Data

Last month, the DC Department of Transportation released a trove of collision data.  The DC DOT and Metropolitan Police Department have a history of sharing information about collisions in the District of Columbia, often through periodic reports.  But this data release is different in that provides information about individual crash report incidents taken by Metropolitan Police Department officers.  So instead of a statistical summary, what you see is individual collisions mapped onto the streets of DC, one collision at a time.

It’s more than a little mesmerizing when the site first loads.  Visually, you seen the map of DC freckled with numerous dots (you must zoom in on the map a ways before you’ll see the individual dots).  Presumably, each dot represents a reported collision for which an MPD officer took a police report (called a PD-10).  Clicking on an individual dot opens up a window that lists some numbers that correspond to the individual crash.  Along busy arterials and traffic circles downtown, the collisions are so numerous they can merge into solid lines in some places.  With more than 20,000 collisions reported per year in the District, you’ll see plenty of data for pretty much  every area and neighborhood you look at.

But after the initial amazement wears off, you’ll probably find that there is not a lot of explanation about what is actually contained on the website.  For example, it doesn’t clearly state what time period the data covers.  After browsing through some of the data tables, it appears that the overwhelming majority of the collisions occurred between 2008 and the present.  There is a bit of explanation about what each field means, some of them make no sense to viewers.

Overall, it appears the website is geared more toward people that routinely handle this type of data as opposed to members of the public that are interested in learning more about crashes in the District of Columbia.  There is not an easy way to sort or visualize the data on the map.  For example, despite my best efforts, I cannot figure out how to limit the map to view only bicycle-related collisions.  Without a way to easily interact with this tool, it’s hard to come up with an idea of how this tool can be used to inform decisions about how DC handles collisions.  With a standardized data set such as this, it seems ripe for a better interface with better ways to visualize the data.

For those that want to dig a little deeper, the site allows you to download the entire dataset.  After I did this and opened it up in Excel, I was faced with an impenetrable wall of numbers 49 columns wide and 216,729 lines deep.  If I have a few hours sometime in the future, I’ll dig in a little and report back.