Impacts of new public transportation stops on bike-sharing demand: A counterfactual analysis using SeoulBike data

Bike-sharing offers a flexible and accessible mode of transportation, particularly for short-distance travel before or after using public transit as a “first-mile” or “last-mile” solution. As a fundamental component of integrated urban mobility systems, bike-sharing enhances multimodal connectivity and complements pedestrian accessibility. Given their integration with transit systems, the installation of new public transportation stops can influence bike-sharing demand in complex ways. This study quantifies these impacts through counterfactual analysis using the Bayesian Structural Time Series model. Rental and return data from SeoulBike, covering 2,843 stations from September 2015 to December 2025, are analyzed to compare actual and predicted demand under a hypothetical scenario in which new transit stops had not been installed. The resulting impacts are classified into three categories: relative increase, relative decrease, and no significant change. Although overall trends indicate a general increase in demand, certain locations exhibit decreased usage, suggesting that the complementary or substitutive relationship between transit and bike-sharing varies across local contexts. A supplementary analysis employs geographically weighted regression to quantitatively assess how land use characteristics shape variations in bike-sharing demand. The model captures spatial heterogeneity by examining the influence of commercial, residential, industrial, and green areas, as well as the density of nearby bus and subway stops. The findings provide actionable insights for optimizing bike-sharing station placement and operational strategies tailored to local environmental factors, thereby supporting urban transportation planning.