Modelling Annual Bike Share Ridership at Hubs with Bike Share Expansion in Mind

Abstract

Public bike share systems have been recognized as an effective way to promote active and sustainable public transportation. With the health benefits of bike share becoming better understood, North American cities have continued to invest in cycling infrastructure and impose new policies to not only encourage the usage of bike share systems but also expand their operations to new cities. The city of Hamilton, Ontario, implemented its own bike share system in March 2015. Using the system’s global positioning system (GPS) data for annually aggregated trip departures, arrivals, and totals in 2017, this research explores various environment factors that have an impact on users’ bike share usage at hub level. Nine predictive linear regression models were developed for three different scenarios depending on the type of hubs and members for trip departures, arrivals, and totals. In terms of variance explained across the core service area, the models suggested the main factors that attract users were distance to McMaster University and the number of racks available at hubs. Furthermore, the working population and distance to the Central Business District and the closest bike lane in the immediate vicinity (200 m buffer) also played important roles as contributing factors. Based on the primary predictors, this research takes one step further and estimates potential trips at candidate sites to inform future expansion of public bike share system. The candidate locations were created on appropriate land uses by applying a continuous surface of regularly shaped cells, a hexagonal tessellation, on the area of interest. The estimated potential usage at candidate sites demonstrated that the east part of the city should be targeted for future bike share expansion.