Sharing behavior in ride-hailing trips: A machine learning inference approach

Ride-hailing is rapidly changing urban and personal transportation. Ride sharing or pooling important to mitigate negative externalities of ride-hailing such as increased congestion environmental impacts. However, there lacks empirical evidence on what affect trip-level behavior in ride-hailing. Using a novel dataset from all trips Chicago 2019, we show that the willingness riders request shared ride has monotonically decreased 27.0% 12.8% throughout year, while trip volume mileage have remained statistically unchanged. We find decline preference due an per-mile costs shifting shorter solo. ensemble machine learning models, travel impedance variables (trip cost, distance, duration) collectively contribute 95% 91% predictive power determining whether requested share it successfully shared, respectively. Spatial temporal attributes, sociodemographic, built environment, transit supply do not entail at level presence these variables. This implies pricing signals are most effective encourage their rides. Our findings shed light can help devise strategies increase ride-hailing, especially demand recovers pandemic.