Sun glare is a recognized hazard reducing visibility, yet quantitative assessments of its dynamic patterns and impacts on traffic safety remain scarce in urban planning frameworks. This study develops a novel framework that leverages large-scale street view imagery to quantify sun glare in complex urban environments and systematically assess its impact on traffic crashes. Analysis in Houston reveals significant spatiotemporal heterogeneity in glare exposure, which is notably intense along east–west roads and during winter. Confirmed significant by GAM, the inclusion of sun glare increased the crash frequency model’s R 2 from 0.82 to 0.88, accounting for 19.58% of feature importance via SHAP. A distinct threshold effect is identified, with risk peaking at 800 annual hours and modulated by road hierarchy and speed limits. Further spatial clustering analysis highlights distinct hotspots of glare-related crashes, particularly at complex intersections and along corridors vulnerable to glare exposure. These findings provide an empirical foundation for integrating glare mitigation into urban design to foster safer and more resilient urban environments. • A GSV-based framework quantifies sun glare in the urban streetscape. • Glare maps reveal strong spatiotemporal heterogeneity across the road network. • Glare inclusion significantly enhances crash frequency model accuracy. • SHAP shows nonlinear glare-crash risk interacting modulated by street features. • Spatial clustering identifies varied glare risk zones for targeted mitigation.
Yang et al. (Fri,) studied this question.