Unveiling pandemic-driven mobility shifts: A S-GTWR analysis of bike-sharing and taxi systems in Washington, D.C

Abstract

A comprehensive analysis of urban transportation systems is essential for effective planning and management. As a representative combination of public and private mobility services, bike-sharing and taxi systems have undergone dynamic changes, particularly during public health crises. This study employs a semi-parametric Geographically and Temporally Weighted Regression (S-GTWR) model to quantitatively evaluate the impacts of socio-demographic, land use, traffic service, and weather-related factors on bike-sharing and taxi ridership in Washington, D.C., throughout the COVID-19 pandemic and subsequent recovery stages. Utilizing census block group-level data, the findings reveal that bike-sharing usage rebounded to near pre-pandemic levels by 2021–2022, whereas taxi ridership remained at approximately 30 % of its pre-pandemic volume. This disparity highlights significant shifts in mobility behavior. Among several models tested, including OLS, GWR, TWR, and GTWR, the S-GTWR model demonstrated superior performance and was selected for spatiotemporal pattern analysis. The model effectively captured dynamic changes in influencing factors by differentiating between trip origins and destinations, thereby offering valuable insights for policymaking. Notably, the variable representing households without vehicles (AUO) was negatively associated with pre-pandemic trip volume, suggesting a behavioral shift during the pandemic from public transit toward alternative modes like bike-sharing and taxis. These results underscore the importance of targeted mobility strategies in response to evolving travel behaviors. The findings provide actionable insights for urban planners and transportation operators to optimize mobility services and enhance urban resilience during public health crises.