Comparisons of built environment effects on the integration of dockless bike-sharing and metro transit across commuting scenarios: A user ID-based clustering identification method

• A user ID-based clustering method is employed to identify integrated trips. • Built environment effects on the integration in commuting scenarios are compared. • Integrated trips are more frequently observed in the morning for commuting purposes. • Population density shows an inverted U-shape in the metro to workplace scenario. • Mixed land use solely stimulates integration in the metro to residence scenario. Dockless bike-sharing is an effective strategy for addressing the first- and last-mile travel problem in connecting with metro systems. Most existing studies simplified the estimation of transfer behaviors and considered integrated trips as a single category when analyzing the impact of the built environment. This increases the possibility of misestimating integrated trips and instability of built environment effects. Drawing on the bike-sharing project in Nanjing, this study employs a spatio-temporal clustering method to categorize integrated trips into four commuting scenarios based on the relationship between commuters’ residence/workplace locations and metro stations. Gradient Boosting Regression Tree models are subsequently constructed to explore the varying influence of the built environment on integration in these scenarios. The results indicate that:(1) Compared to the evening, integrated trips are more frequently observed in the morning for commuting purposes. (2) Over 30 % of integrated trips are situated beyond 50 m from metro entrances, indicating resistance to the combination of these two modes. (3) The impact of population density shows an inverted U-shape in the metro to workplace scenario. (4) shopping facilities and mixed land use solely exert a stimulating effect in the metro to residence scenario. (5) Both areas in proximity to the city center and situated more than 16 km away exhibit lower integrated trips. These findings provide a better understanding of the specific scenarios and ranges in which built environment variables influence the integration of dockless bike-sharing and metro systems.