Resumen
Although many studies have explored the relationship between the built environment and metro ridership, the literature offers limited evidence on the nonlinear effect of origin and destination built environments on station-to-station ridership. Using data from Chongqing, this study uses the gradient boosting decision trees (GBDT) model to explore the nonlinear impact of origin and destination built environments on metro ridership. The research results show that the built environment at the origin has a greater impact on metro ridership than the built environment at the destination. All the independent variables examined have complex nonlinear effects and threshold effects on metro ridership. The distance to the city center, the number of companies, and the building volume rate have a greater positive effect on metro ridership, both at the origin and at the destination. The research results provide suggestions for optimizing the built environment around metro stations.