There is an interaction between the land use and traffic congestion. In order to quantify the non-stationary impact of the land use on traffic congestion, this paper using the ride-hailing data in Chengdu which published by the GAIA Open Dataset, applies the GWR model to reveal this impact. We find that tested by Sigma, AICc, Adjusted R2, the GWR model was superior to the OLS model in analyzing the impact of the built environment on traffic congestion. And spatial heterogeneity both exists in land use distribution and traffic congestion distribution. On the selected time period, catering points in suburban areas have little impact on traffic congestion, while in the lively region it have a palliative effect on traffic congestion. The more concentrated the distribution of companies and residential areas are, the more serious the traffic congestion is, and these effects are more significant in the suburbs.

CICTP