Analyzing Road Traffic Incident Hotspots Using Cluster Analysis in Thanh Hoa Province of Vietnam

Abstract

This study uses remote sensing images combined with Geographic Information Systems statistical approach - spatial autocorrelation analysis to identify traffic incident hotspot locations. It evaluates the statistical significance of hotspot clusters to support forecasting road traffic incidents applied for Thanh Hoa province of Vietnam with traffic accident data during four years from 2020 to 2023. Based on spatial analysis techniques, including severity index and Getis-Ord Gi* statistical analysis with inverse distance weighting interpolation, hotspot clusters are identified and sorted by rank.  Applying spatial autocorrelation analysis has important implications in enumerating hotspots in sequence. The analysis results identified 64 hotspots, of which the 32 dangerous incident points are in Thanh Hoa province's transportation system. The results demonstrate the effectiveness of using Getis-Ord Gi* spatial statistical techniques to identify traffic incident hotspot locations and evaluate the statistical significance of hotspot clusters to support forecasting road traffic incidents in Thanh Hoa province.