For new rail transit stations, it is easy to deviate from the forecast of passenger traffic, resulting in the site facilities not meeting the passenger flow demand or excess facilities. Taking the three lines of Xi'an subway as an example, the article takes 800m as the radiation range of the site, and uses the geographically weighted regression model to study the relationship between the passenger flow and the spatial attributes (population density, employment). Finally, it is find that: (1) the influence of spatial attributes on passenger flow generation rate is: employment > population density (2) geographically weighted regression model can get better regression effect, population density and employment relationship with subway passenger flow Positive correlation, showing a 1% increase in the two, the total passenger flow of the subway station increased by 0.47% and 0.53% respectively; (3) In the region with small population density and small employment, the relationship between passenger flow and the two is more sensitive. According to the population density within 800m of the built subway station and the employment position, the passenger traffic can be predicted to provide theoretical support for the service level of the new site.

CICTP