Effects of station-area built environment on urban rail transit passenger flow

Abstract

The spatiotemporal mismatch between urban rail transit (URT) stations and surrounding urban functions challenges operational efficiency worldwide. Existing research predominantly focuses on average daily ridership, offering limited insights into how built environment influences vary across station types, time periods, and spatial scales. This study develops an integrated framework combining station classification with multiscale geographically weighted regression (MGWR) to investigate the mechanisms linking built environment to passenger flows. Using Changsha, China as a case study, stations are classified into residential, employment, and mixed-use types via K-Means clustering. The MGWR model is compared with OLS model and GWR model. Results reveal three key findings: built environment effects exhibit spatiotemporal heterogeneity, functional-type dependence, and spatial scale variability; factors such as office density have divergent effects across station types and between morning and evening rush hours. The framework effectively identifies local and global operational scales, supporting a paradigm shift towards station-specific, refined governance strategies.