Analyzing locations prone to causing road crash injuries and fatalities in Saudi Arabia

Abstract. This study proposes a systematic approach to analyzing crash data and identifying high-crash areas in Saudi Arabia’s intercity road network. The approach employs a five-stage process involving data preparation, cluster identification using density-based clustering algorithm DBSCAN, predictive modeling for fatalities and injuries, data aggregation, and cluster ranking. This research’s results reveal that nearly a quarter of crashes are concentrated within clusters. The predictive model confirms the expected relationship between vehicle type, time of day, and crash severity. By quantifying excess fatalities and injuries in these clusters, safety analysts can prioritize locations that should receive targeted interventions. This research offers a practical framework for enhancing road safety. By automating the proposed approach and providing clear visualizations, safety analysts can efficiently identify high-risk areas and implement data-driven countermeasures. This study represents a significant step toward reducing crashes and their consequences.