This study comprehensively analyzes pedestrian-runover crash density (the number of crashes per square kilometer of district area) in Riyadh’s school zones, employing advanced Artificial Intelligence (AI) techniques, including Multiple Linear Regression (MLR) and Machine Learning (ML), to enhance urban efficiency, quality of life, and resilience. Data were collected from 884 school zones distributed across Riyadh, encompassing diverse infrastructural, socioeconomic, and demographic contexts. The optimized MLR model identified significant predictors, including district road lengths, average crash severity (EPDO), average income, population density, and transit stop availability, which collectively explained approximately 65% of the crash-density variability. The Random Forest ML model further improved predictive accuracy (R² ≈ 0.88), revealing complex, nonlinear interactions among key variables, including traffic volume, speed limits, lane counts, crosswalk availability, and student population. Integrating traditional regression with cutting-edge ML methodologies, this research provides actionable insights for policymakers, urban planners, and engineers, enabling targeted, data-driven interventions to enhance pedestrian safety and promote sustainable, smart urban mobility in Riyadh’s school zones.
