Spatial Analysis of Road Traffic Accident Hotspots in Jega, Nigeria: A Comparative Study of Kernel Density Estimation and Geographically Weighted Regression

Road traffic accidents remain a critical public safety challenge in rapidly urbanizing regions of sub-Saharan Africa, where heterogeneous road infrastructure and high population density exacerbate risk. This study applies Kernel Density Estimation (KDE) and Geographically Weighted Regression (GWR) to analyze spatial patterns of road traffic accidents across Jega Local Government Area, Kebbi State, Nigeria, using fifty georeferenced primary data points collected through Global Positioning System surveys and manual traffic counts. The KDE analysis identified optimal bandwidth of 175 meters with a Prediction Accuracy Index (PAI) of 3.50 at the 85th percentile threshold, indicating strong spatial clustering of accidents. Spatial autocorrelation analysis revealed significant clustering (Moran's I = 0.312, p < 0.05). The GWR model demonstrated strong explanatory power with global R² of 0.72 and AICc of 420.35. Local R² values exhibited substantial spatial variation (range: 0.20–0.95), highlighting the importance of localized analysis. Cross-validation results (RMSE = 3.45, MAE = 2.12, R² = 0.65) confirmed predictive robustness. The integrated geospatial framework identified distinct high-risk corridors, with Gada (8 accidents), Garkar Ando (5 accidents), and Gobirawa (5 accidents) emerging as critical hotspots requiring immediate intervention. This research provides a validated geostatistical framework for micro-scale road safety planning in Nigerian cities.