Since automatic fare collection data could not to accurately reflect the influencing factors around the station. This paper proposes a method to analyze the influence factors of passenger main distribution areas around subway stations by 270,000 Points of Interest data and 180 million cell phone data in Chongqing. Firstly, based on the cell phone data triggered in entering and exiting subway stations, ST-DBSCAN is used for clustering the main passenger distribution areas as Eps=500, Minpts=4, where 85% of cell phone data recorded by each base station can be clustered well. Then, Structural Equation Model is established to analyze the influencing factors between passenger flow distribution and built environment. At last, the standardized estimation parameter of POIs is 0.725 and the public transportation facility is 0.577 in latent variables. Among observed variables, the most important influencing factor is restaurant, which is 1.007, and the least is parking lots, which is 0.376.

CICTP