How do built environment characteristics influence bus use patterns across neighbourhood types in Beijing? A machine-learning analysis

Abstract

There is a growing interest in the link between the built environment and travel behaviour in Chinese cities. However, the extent to which the built environment impacts transit use across various neighbourhoods under different planning and development strategies remains ambiguous. This study focuses on Beijing and identifies four distinct neighbourhood types: inner-city, transient, clustered suburban, and dispersed suburban. Utilising smart card data in conjunction with a machine-learning technique—XGBoost—we comprehensively explore the intricate connections between built environment characteristics and bus usage at the neighbourhood level. The results indicate that regional location, bus route availability, and density are more important in affecting bus trip frequency and duration. Meanwhile, the importance of these variables and associations with bus use differ substantially across different types of neighbourhoods. Notably, opposite effects are observed for road density and distance to the nearest subcentre in inner-city, transient, and suburban neighbourhoods. Based on these findings, there are important implications for neighbourhood planning to support bus usage. By considering the varying factors that impact bus use in different neighbourhoods, planners can develop more effective strategies to encourage and facilitate bus travel.