Probabilistic Market Segmentation of Transit Users in Vancouver, Canada, During the COVID-19 Pandemic

Before the COVID-19 pandemic, Metro Vancouver was the North American leader in transit ridership growth, with its most popular routes being severely crowded during peak travel times. As the region gradually recovers from the effects of the pandemic, it is expected that the crowding will also return to the transit system and will have to be tackled. This study performed a probabilistic classification of transit users in Metro Vancouver that could be used to develop a set of policy interventions aimed at distributing the peak hour use of transit services to other times, or less crowded routes. Principal Component Analysis was employed to explore the underlying relationships between the attitudinal indicators that informed the specification of the classification model based on the Hybrid Discrete Choice framework. Such socio-demographic factors as being a female, of working age or a senior, having children, bachelor’s degree or higher, having low income, or traveling during morning peak hours were found to influence the latent variables in the classification. The final model produced six probabilistic classes based on the estimates for two latent variables that accounted for respondents’ concerns regarding crowding and safety, as well as personal flexibility to travel to and from work via public transit. Based on the results, a policy framework is developed that suggest that Metro Vancouver’s transit agency might already be affecting the travel choices of those riders who are most concerned and flexible through the provision of information on crowding levels. On the other hand, to affect the choices of those riders who are least flexible, it is recommended that the agency develop partnerships with large regional employers