<p>This research investigates the spatial and temporal relationship between taxi and public transit usage in order to probe the role of taxi in facilitating public transit and enhancing transport sustainability. By geographically aggregating large volumes of high-granularity taxi GPS data from New York City into block-group level metrics and integrating with a variety of built environment features and socio-economic factors, we apply spatial econometrics models. We find that public transit coverage (subway and bus) in a neighborhood has positive impacts on the number of taxi drop-offs during morning peak hours and taxi pick-ups during evening peak hours, both in that neighborhood locally and in its surrounding neighborhoods. These positive direct and spillover effects are statistically significant after controlling for employment density. This research has important policy implications on how the growing on-demand ride-sourcing services should be regulated to reinforce their complementary role to public transit, enhance transit ridership, and promote transport sustainability.</p>