Ride-hailing services have become major contributors to urban transportation systems and may influence traditional modes such as metro systems. While previous studies have explored how ride-hailing services either complement or substitute public transportation, limited research has investigated the nonlinear relationship between ride-hailing and metro ridership across different urban land uses. To fill this research gap, this study examined the nonlinear relationship between ride-hailing and metro ridership among six specific areas (commercial, green space, office, university, healthcare, and transportation hub) from global and local perspectives. We employed the light gradient boosting machine (LightGBM) model to predict metro ridership and capture the nonlinear effects of ride-hailing ridership. The Shapley additive explanations (SHAP) method was then utilized to interpret the model predictions. Using data from the case of Chicago, the feature importance analysis demonstrated that ride-hailing ridership had an important effect on predicting metro ridership. Additionally, the highest relative importance of ride-hailing on metro ridership was observed in commercial areas on Saturday. In contrast, ride-hailing’s influence was most pronounced on Friday for other land-use types. Furthermore, the results uncovered varying degrees of nonlinearity in the relationship between ride-hailing and metro ridership across commercial, green space, office, university, and healthcare areas.
Exploring Spatiotemporally Varying and Nonlinear Relationships between Ride-Hailing and Metro Ridership: A Case Study in Chicago
2025-10-01
Article (Journal)
Electronic Resource
English
Exploring the correlation between ride-hailing and multimodal transit ridership in toronto
Online Contents | 2021
|How Built Environment Impacts Online Car-Hailing Ridership
Transportation Research Record | 2020
|European Patent Office | 2025
|RIDE-HAILING SYSTEM, RIDE-HAILING METHOD, AND COMPUTER PROGRAM
European Patent Office | 2024
|