Traffic accidents

combining open data sources to investigate influences on traffic accidents

Scroll down for information


Traffic Accidents


February 2015


Combining several open data sets to look for trends


Illustrator, R


Accidents: Bestand geRegistreerde Ongevallen Nederland
Rain & Snow: KNMI
Daylight hours:


This infographic gives a visual way for the viewer to compare the number of traffic accidents in the Netherlands during 2013 to weather patterns, school holidays and daylight hours.


At work I was discussing with some colleagues what external datasets we could use to create a prediction model of burglaries. And suddenyl I thought that perhaps the same datasets could also say something about the number of traffic accidents. I had wanted to make a circular visualization for a while and this seemed like a good opportunity.

I found all datasets; the traffic accidents, the amount of rain, the snow days (this was actually tucked away in PDF reports), holidays and the number of daylight hours (had to webscrape this). I created the three line charts (daylight hours, traffic accidents & LOESS curve and rain fall) in R and then imported these graphs in Illustrator where I bend them into a circle and added all the colors, icons and annotations. For the latter I searched the news for that particular date to see if anything  could be found relating to traffic jams and such.

In the end I do see the correlation between holidays or daylight hours and the number of traffic jams. From the annotations I also know that snow during rush hour gives the worst days but not all snow week days resulted in many accidents. Rain shows even less correlation. I guess using the daily average of one station in the center of the Netherlands is too generalizing, even if the Netherlands is already rather small.

You can read more about the data gathering and design in my blog.