Using Global Positioning System Big Data to Define a Bikeshare User Typology and Model On-/Off-Road Bicycle Facility Preferences

Abstract

Bikeshare programs have grown exponentially over the last decade and have been found to be influential in increasing the nonmotorized mode share. With shared microbility programs proving to be a key component of modern urban environments and in first- and last-mile connections to transit, there is a need to better understand users and their facility preferences. This study utilizes 3 years’ worth of data from a campus bikeshare program in Auburn, AL, that includes over 100,000 trips and almost 9,000 users to develop a bikeshare user typology. Whereas past research developing cyclist groupings has been based on self-reported cyclist comfort and confidence levels, this study utilized actual ridership data from the bikeshare program to create new typologies specific to bikeshare. Using a K-means cluster analysis, this study proposes four bikeshare user types: unsupported users, aspirational users, flexible users, and purposeful users. These users have varying travel motivations, needs, and patterns. Their choice between an on-road and off-road facility option was analyzed at nine specific locations displaying different roadway and physical conditions. A binary logistic regression analysis was performed to determine which variables are influential in a bikeshare user’s choice between riding on an on-road or off-road facility. The modeling results found two typology memberships were a significant predictor of bicycle facility preference, keeping all other factors constant. Specifically, flexible and purposeful users had stronger preferences for off-road bicycle facilities. Additionally, the physical conditions were found to be significant with users choosing off-road options away from motor vehicle traffic as the roadway environment becomes more intimidating.