How does the urban built environment affect dockless bikesharing-metro integration cycling? Analysis from a nonlinear comprehensive perspective

The multi -modal combination of "dockless bikesharing (DBs) + metro" can effectively addresses the "last mile" challenge, and its importance to improve the quality of metro integration and residents ' commuting efficiency is becoming increasingly prominent. As one of the important determining factors of transportation, the urban built environment directly affects the integration travel behavior of residents. To better understand this issue, it is necessary to conduct in-depth research on the influence mechanism of urban built environment on DBs-metro integration cycling (DBsMIC), and reveal its complex nonlinear characteristics and spatial heterogeneity. Therefore, taking Shenzhen as a case study, this study first introduced the OLS regression model to detect the significant built environmental impact factors of DBsMIC. Secondly, the random forest regression model was used to analyze its nonlinear influence. Finally, aiming at the problem of spatial non-stationarity of influencing effects, the spatial heterogeneity is analyzed from both global and local perspectives based on the MGWR model. The results show that the significant impact indices of the built environment of the integration cycling volume vary in different time periods. Overall, seven factors such as road and residential density have a significant impact on DBsMIC. Among them, the job-housing balance and the bus line density have a negative impact, while the rest have a positive impact. Meanwhile, the nonlinear effect of the built environment on DBsMIC is obvious, and there is an obvious marginal effect. In addition, the nonlinear impact of the built environment on DBsMIC in different time periods has spatial heterogeneity and spatial marginal effects. For example, the influence of road density shows a global positive correlation effect in multiple time periods, and the promotion effect increases from southwest to northeast. This study can provide auxiliary decision-making support for the construction of an efficient, green and sustainable urban transportation system according to local conditions.