I’m currently working on an analysis of a huge database with geocoded bicycle accidents for the urban area of Salzburg (Austria). The plan is to extract new information through an explicitly spatio-temporal analysis at multiple scales. This means that not only the factors which potentially contribute to a bicycle crash are considered but their pattern in space and time.
One hypothesis is, that different risk factors contribute to the chance of an accident depending on location and time. For example, we do know from the data that alcohol consumption plays a significant role from late afternoon to the early morning depending on the season; but the effect is limited to certain locations in the city center and along the Salzach river. This is where most bars are located and where (mainly) teenagers drink at the promenade. In a follow-up post I’m going to publish a couple of nice charts which show the patterns such as the described one.
Of course I’m curious about what has been done in this field so far. There is a lot of research in medicine (they are primarily interested in the injuries of course), traffic engineering and planning. GIS is not very prominent in this respect. But what surprised me was the fact, that the number of publications which deal with bicycle accidents and GIS are increasing exponentially. Just the same like the more general research of bicycle accidents. One can say, that the explicit geographical view on the topic is becoming more important.
Of course a simple Google Scholar search has several drawbacks and biases (double count, no completeness etc.). But for a first approximation it is quite useful. If you’ve conducted any kind of research in this area or you have suggestions for further readings and sources, please let me know!