Mapping Cyclist Activity and Injury Risk in a Network Combining Smartphone GPS Data and Bicycle Counts

Author(s): Jillian Strauss, Ph.D. (Candidate), Luis F. Miranda-Moreno, Ph.D., Patrick Morency
Student Paper Competition: 2nd Place

Slidedeck Presentation:

1A - .Strauss


In recent years, research has been carried to identify environmental risk factors and map injury risk for cyclists. These tasks require three main sources of data: geocoded injury data, geometric design and built environment characteristics as well as exposure measures, also referred to as motor-vehicle and bicycle flows, volumes or activity. Bicycle flow data on each facility and network element is an essential component in the calculation of cyclist injury rates (also referred to as risk). Bicycle flows are required to identify routes and corridors with high injury risk or with high bicycle activity. This knowledge will serve as vital information for cities wishing to implement appropriate cycling infrastructure. The main objectives of this study are to estimate and map bicycle volumes, injuries and risk throughout the entire network of road segments and intersections on the island of Montreal, combining smartphone GPS traces and cyclist counts (manual and automatic, short-term (hours) and long-term (months and years)) to then validate the use of GPS data as a potential source of cyclist exposure data. Bayesian methods are applied to the GPS data to map cyclist injuries and risk throughout the entire island of Montreal. Among other results, cyclist risk is greatest outside the central neighbourhoods and where bicycle infrastructure is not present and much greater at intersections than along segments. This study validates the use of GPS data as a new and reliable source of bicycle flow estimation, useful in a variety of safety analyses carried out at the entire urban network level.