Pedal cycle incident data in KML

The innovation blog just released a data set of the “Pedal Cycle Accident Locations” for 2005-2007, broken down by year, in XLS.

I had early access to this at Rewired State and converted the Easting/Northings to Longitude/Latitudes through Multimap’s conversion API, and then from CSV into KML. Get the files for 2005, 2006 and 2007 and have a poke around.

It’s a really interesting data set, used wisely. There are around 16000 incidents a year (let’s not call them accidents — often one of the parties is to blame), reported to the police. The definition is:

Accident: Involves personal injury occurring on the public highway (including footways) in which at least one road vehicle or a vehicle in collision with a pedestrian is involved and which becomes known to the police within 30 days of its occurrence. The vehicle need not be moving, and accidents involving stationary vehicles and pedestrians or users are included. One accident may give rise to several casualties. “Damage-only” accidents are not included.

The locations seem to be relatively accurate (in London, at least), overlaying the road network close enough that you can pinpoint roads and even junctions.

I’d like to snap the points to the nearest road using OpenStreetMap, divide by road length and get an idea of the most dangerous roads and who is responsible for them. Alas, this probably involves me seriously levelling up in Postgres/PostGIS, which I haven’t found the time to do yet.

You could also use the data as a weighted input to a route mapper, helping to prioritise safer roads.

It’s important that it’s used positively. Cycling is safe, and more cycling makes it safer for everyone.