Data-Driven Prioritization of Road Traffic Black Spots in Upper Egypt Using GIS, Machine Learning, and Economic Risk Analysis

Road traffic accidents (RTAs) continue a critical health and economic challenge in Egypt, particularly in Upper Egypt where rapid population growth and infrastructure limitations intensify road risks. Despite national safety initiatives, traditional black spot identification methods still rely mainly on accident frequency, overlooking severity, economic impact, and spatial dynamics. This study develops an integrated, data-driven framework to proactively identify and prioritize hazardous road segments by combining crash severity scoring, economic cost estimation, GIS spatial analysis, and predictive machine learning models. The research applies the framework to the El-Minia–Beni-Suef passageway using both historical crash records (2010/2014) and recent datasets (2020/2024). Critical crash rates are estimated, kernel density estimation (KDE) is employed to detect spatial clusters, and optimized random gradient boosting algorithms are trained to identify current and emerging high-risk segments. Comparative evaluation shows that the proposed framework achieves substantially higher predictive accuracy than frequency-based approaches, offering more reliable insights into hidden accident patterns. The results determine the value of participating in ML with spatial and economic diagnostics to support proactive resource allocation and evidence-based safety interventions. The framework directly supports Egypt’s Vision 2030 and aligns with the UN Sustainable Advance Goal (SDG) 3.6 of reducing traffic-related deaths and injuries. Additionally, the approach is scalable and transferable, providing a practical model for other mounting countries challenging similar road safety challenges.