Increasing concerns about the safety of cyclists request for more research in how infrastructure can be improved to increase cycling safety. Literature in the prediction of bicycle crash risk however is very limited. Also the amount of available (exposure) data was often very limited. The aim of this study was to provide a better understanding of how infrastructural factors contribute to the safety of cyclists and how cycling safety can be improved. By combining rich exposure data from the FietsMonitor of Witteveen+Bos with crash data and infrastructural data, safety performance functions were developed for both road sections and intersections. Negative binomial generalised linear regression modelling (GLM) was used to find the relationships between bicycle crash risk and infrastructural risk factors. Most relationships that were found were as expected. Separate bicycle facilities, the presence of speed humps, closed pavements and an increased bicycle lane width were all found to be related to lower bicycle crash risks. Regarding intersections, the amount of traffic flows crossing the intersection did significantly increase bicycle crash risk. Moreover, the results showed a higher bicycle crash risk for roundabouts compared to other intersection types, which was unexpected. Finally, street light was found to be significantly positively related to bicycle crash risk for both road sections and intersections. As the safety performance functions are incomplete, further research should look into other explanatory infrastructural factors. Furthermore, it is recommended to extend the data to other cities to validate the results.
The impact of infrastructure design on cycling safety
Master thesis Delft University of Technology