This study's goal is to identify the geographic hotspots of traffic accidents in Tunisia's coastal and inland areas using the spatial autocorrelation approach, with particular emphasis on the Moran and Getis-Ord indices. The approach used also allows adapting to the spatial structure observed locally, through local measures of spatial autocorrelation. The results show a fairly significant disparity in the concentration of these points between regions, types of road, geographic locations, and environments (urban, rural). In coastal areas, the black spots are, also distributed in the governorate of Nabeul, Sfax and Sousse along the national and regional roads. However, in the inland areas, the black spots are, distributed and found in a more abundant way in the governorates of Kef, Beja and Kairouan all along the rural roads. The precision brought by our results could help the public decision-maker as on accident prevention, particularly given the significant growth in the number of cars in Tunisia.

Keywords

Traffic accidents, black spots, spatial autocorrelation, Moran index, Getis-Ord index.