Hierarchical Linear Model for Investigating Effect of Built Environment on Bus Transit

Because proper urban planning and design can improve transit sharing, the impact of the built environment on transit use has attracted the attention of many researchers. Factors with different geographical scales such as individual characteristics, neighborhood built environment, and regional features can influence transit use. However, the nested structure of geographical scales has not been considered in previous research studies. By considering the nested geographic features, a bilevel hierarchical linear model (HLM) is established in this study to explore the combined influence of the surrounding built-environment factors at the neighborhood level and the socioeconomic variables at the regional level on bus transit trip rates. The model is implemented based on land-use data and bus smart card data of approximately 3 million records of 9 districts and more than 1,600 bus stops in Wuhan, China. The results present evident differences in the effect of the built environment on transit use in different spatial districts. The population and public transportation investment at the regional level are important factors leading to these differences and influence transit use. Based on the geographically nested features in the data, the proposed model reveals the impacts of the built environment on transit use and investigates the causality of the reported discrepancies. Furthermore, combined with the model results, optimization strategies for improving bus transit use are proposed for a region in Wuhan. The proposed method offers theoretical and technical support for transit-oriented urban development.