Evaluating the impact of bicycle behaviour on safety

Bicycling count data (exposure) are difficult to collect but are vital to understanding the safety of different routes. Crowdsourced exposure data collected through fitness apps such as Strava Metro can provide bicycling exposure data with better spatial and temporal coverage than manual or automatic counts. Efforts are underway in cities throughout Canada and the US to use Strava data to estimate both the number of bicyclists at each segment in road network, as well as to characterize the type of exposure based on patterns in the data through time (e.g. commuting versus recreational bicycling). The student (Mr. Branion-Calles) will utilize data from Ottawa, Canada as a case study to develop a statistical model of bicycling crash risk at the network level. The integration of both overall exposure and the type of exposure, into crash risk models can provide insight into the drivers of differences in safety between physical route characteristics. Network level crash risk models for bicycling are rare, and none have included temporal characterizations of exposure, rather they assume all exposure is equal. The expected outcome is one manuscript for peer-review in a scientific journal.

Faculty Supervisor:

Meghan Winters

Student:

Partner:

Arizona State University

Discipline:

Sociology

Sector:

Transportation (excluding aerospace); Health and Related Sciences & Technology

University:

Simon Fraser University

Program: