Understanding the relationship between built environment and emerging mobility services: Evidence from designated driving and ride-hailing in Beijing

Emerging mobility services are reshaping urban transportation systems. Although the associations between the built environment and ride-hailing demand have been extensively studied, how the built environment nonlinearly associates with designated driving demand and how these associations differ from ride-hailing remain systematically underexplored. This study develops an analytical framework integrating adaptive spatial unit delineation with spatially dependent Gradient Boosting with Gaussian Process (GPBoost). Based on trip data from Beijing, China, we systematically identify these nonlinear associations and differentiated response mechanisms using SHAP (SHapley Additive exPlanations) methods. Results demonstrate that adaptive triangulation with a minimum edge length of 1000 m combined with the GPBoost model effectively accommodates the spatial unevenness of demand data distribution and significantly improves model fit and predictive accuracy. The two services exhibit systematically different nonlinear association patterns with built environment factors: designated driving demand is driven by specific scenarios, displaying pronounced temporal dependency and threshold saturation characteristics, whereas ride-hailing demand stems from diverse travel purposes, exhibiting multi-stage nonlinear responses and all-day effects. Specifically, high-earner density is significant for designated driving only during high-demand periods but significant for ride-hailing throughout the day; bus stop density shows an inverted U-shaped association with ride-hailing; and car maintenance facilities are uniquely associated with designated driving demand. The study reveals how service functional differentiation systematically shapes differentiated demand generation patterns, providing empirical evidence for platform operators' spatial dispatching optimization and urban planning departments' facility configuration.