Abstract
Like many countries, the extent of road safety in Tunisia is considerable. In addition, road accidents cause extensive economic losses. It is in this perspective that the requirement of safety has become a priority value, a goal that applies to everyone, especially in the field of transport. The objective of this research is to examine pedestrian-vehicle collisions (PVC) in coastal regions in Tunisia in term of area of occurrence under peak time periods. To do so, an innovative and integrated two-step approach was applied. Firstly, Ripley's K-function was used to detect the highway segments, where PVC were significantly clustered, dispersed or random. Secondly, Since PVC are distributed over a Network Kernel Density Estimation approach will be used to detect where exactly road accidents aggregates across the study area under peak time periods in order to predict some hourly trend to the PVC distribution and to find whether the detected high-risk locations are subject to temporal fluctuation. To the best of the authors' knowledge, this research is the first to examine road accidents in term of hotspot analysis under peak time periods. The findings of this has significant implications for public decision-maker to enhance road safety situation, especially with the considerable surge in the vehicle fleet in Tunisia.