Geospatial Model for Identifying and Assessing Risk at Hazardous Locations in the Road Network Based on Environmental and Infrastructure Characteristics

This article presents a geospatial model for identifying and assessing the risk of hazardous locations in the road network, developed to predict traffic safety hazards in areas with complex infrastructure where traditional methods, such as the Highway Safety Manual, are insufficient. The objective of the study was to develop a model that classifies road segments into five risk categories based on environmental and infrastructural characteristics, without using accident or traffic volume data. The model accounts for speed limits, road geometry, and the proximity of facilities that generate pedestrian traffic (schools, preschools, stores) and infrastructure elements (crosswalks, intersections). A hybrid approach was used, combining proprietary methods for determining distances from objects: vector-based (geodetic distance), route-based (road graph), and geometric (classification of a road segment’s shape), using QGIS, OpenStreetMap, and custom Python scripts. The results enabled assigning a risk category to each road segment, and validation was performed by comparing them with the locations of actual accidents resulting in serious injuries or fatalities. The developed model for identifying hazardous locations is a scalable tool that supports sensor-network-based area-based speed control systems, infrastructure planning, and safety management in regions with diverse road networks.