Evaluating the impact of rail transit network expansion on travel behavior in Shenzhen, China: A causal analysis across different stages of development

Research regarding the impact of rail transit development on travel behavior has received significant attention in recent decades. Numerous studies have examined the influence of newly established rail lines on travel behavior. However, there is a noticeable lack of studies on how the expansion of a rail transit network alters daily travel behavior. This study addresses this research gap by investigating the travel behavior impacts of the rail transit network expansion in Shenzhen. A machine learning (ML) enhanced difference-in-differences (DID) model is developed to determine the causal effects of enhanced transit accessibility on travel behavior changes. Our findings suggest that enhancements in rail transit accessibility significantly boost rail transit use and decrease travel by private cars and buses. We also discovered that the effects fluctuate at different stages of rail transit network development. These findings bear significant relevance for strategies and policies concerning rail transit development and transport demand management.