GIS-Based Spatial Analysis of Accident Hotspots: A Nigerian Case Study

This study identified high-risk locations (hotspots) using geographic information systems (GIS) and spatial analysis. Five years of accident data (2013-2017) for the Lokoja-Abuja-Kaduna highway in Nigeria were used. The accident concentration analysis was conducted using the mean center analysis and Kernel density estimation method. These locations were further verified using Moran's I statistics (spatial autocorrelation) to determine their clustering with statistical significance. Fishnet polygon and network spatial weight matrix approaches of the Getis-Ord Gi* statistic were used in the hotspot analysis. Hotspots exist for 2013, 2014, and 2017 with a significance level between 95-99%. However, hotspots for 2015 and 2016 have a low significance level and the pattern is random. The spatial autocorrelation analysis of the overall accident locations and the Moran's I statistic showed that the distribution of the accidents on the study route is random. Thus, preventive measures for hotspot locations should be based on a yearly hotspot analysis. The average daily traffic values of 31,270 and 16,303 were obtained for the northbound and southbound directions of the Abaji-Abuja section. The results show that hotspot locations with high confidence levels are at points where there are geometric features.