This is just a short post about the “STATS19” open dataset on where road accidents logged by the police happen. There are many great things about this dataset (apart from it being about injury and death), my favorite being

  1. it’s been made open to the public, enabling some interesting DIY analyses and visualisations such as this one by CycleStreet.net and
  2. every accident is geo-referenced quite accurately, allowing geographers like me to make maps of it and try to understand the causes and effects of this risk.

The reason I’m getting excited about this project right now is that I’ve just started work on a new project with Dr Ian Kellar that seeks to harness it. So I’ve imported it into my favourite stats program R and made some preliminary visualisations including this little map of accident locations in a small area between Leeds and Bradford:

There’s also this, which provides an indication that accidents involving cyclists are most likely to happen at slightly different times than other types of accidents.

Accident times

Both these mini visualisations have made me keen to investigate this data resource further, so watch this space on the project’s page on GitHub. If you are really keen, feel free to fork it and do your own research: I’m intending to do it all with open source tools so the results should be accessible to anyone. Enjoy!