We combined three publicly accessible datasets regarding biking in Greater London to study bike safety. The three datasets were:
1. Bike accidents in the United Kingdom from 2005-2018
2. Annual vehicle and bike traffic for most roads in the United Kingdom from 2000-2021
3. Bike lanes in Greater London
We used these datasets and nearest neighbor algorithms to build a new dataset which contained traffic flow, number of accidents (including severity), and nearby bike safety features for most roads in Greater London. We then performed k-means clustering on this new dataset to determine whether shared bus and bike lanes were as safe as segregated bike lanes.