Effects of built environment on metro ridership at a microscopic scale: a case study of Xi'an, China

Few studies have examined the relationship at the microscopic spatial scale. In this study, multiple sources of data including mobile phone signal data, automatic fare collection system data, geo-information data, and street-view image data are combined to measure metro ridership and built environment at the plot or block scale. The Random Gradient Boosting Decision Tree was used to explore relationship between the built environment and ridership. The results show the following: (1) the relationship between built environment and ridership shows different types of curves. (2) The path distance to the metro station and the visual perception of road space have more significant impacts on ridership than road network density. (3) The location of the grid also affects grid-level metro ridership. The results suggest that planners should consider the locational factors, pay attention to the different effective thresholds of different variables on ridership and the longitudinal landscaping of non-motorized urban roads.

Automated summary

This study examines the relationship between metro ridership and the built environment at a microscopic spatial scale. The results show that the built environment has different types of curves in relation to ridership. Factors such as path distance to the metro station and visual perception of road space have a more significant impact on ridership than road network density. Additionally, the location of grid cells also affects metro ridership. The study suggests that planners should consider locational factors and the longitudinal landscaping of non-motorized urban roads.