How is commute mode choice related to 2D and 3D land use mix? a case study of central Suqian City

Understanding residents’ commute mode choice and its influencing factors is crucial for alleviating traffic congestion. Current research on land use mix has largely focused on 2D space, with limited attention to how 2D and 3D land use mix and their threshold effects differentially influence residents’ commuting behavior. To address this gap, this study empirically examines Suqian, a prefecture-level city in China, by deriving land use mix characteristics from land use, building, and POI data. Then applied the XGBoost model to test and compare the nonlinear impacts of land use mix features on the mode shares of different commute modes. The results indicate that the importance and threshold effects of 2D and 3D land use mix variables differ across commute mode choices, thereby identifying their effective ranges. The study theoretically enriches the land-use and transport interaction framework and practically provides references for land use policies to alleviate traffic congestion.