Identification of Hotspots and Spatial Patterns of Traffic Crashes in Fortaleza (CE) in 2024

Abstract

This study analyzed the spatial distribution of traffic crashes in Fortaleza by integrating three methods: Nearest Neighbor Analysis (NNA), Kernel Density Estimation (KDE), and Local Indicators of Spatial Association (LISA). The results showed that crash events are not randomly distributed but exhibit a strong clustering pattern, as indicated by Nearest Neighbor Index (NNI) values below 1 and significantly negative Z-scores. KDE allowed the identification of high-density corridors, particularly along BR-116 and major arterial roads, while LISA highlighted local spatial structures, revealing persistent High–High clusters and critical outliers located in specific segments of the road network. Findings demonstrated that fatal crashes are concentrated on high-hierarchy, high-speed roads, whereas property-damage-only crashes display a more diffuse pattern associated with densely urbanized areas. Overall, the analysis indicates that traffic crash distribution in Fortaleza is shaped by structural factors related to road configuration, land use, and mobility dynamics, reinforcing the need for localized, spatially oriented interventions. The results confirm the relevance of integrating spatial analytical methods as a support tool for urban planning and road safety management.