Spatio-temporal analysis to identify hazardous road locations for vulnerable road users

According to the World Health Organization, each year approximately 1.3 million people die as a result of road traffic collisions. More than half of all road traffic fatalities are among vulnerable road users which include pedestrians, cyclists, and motorcyclists. The aim of this study is to identify and evaluate methods for the spatio-temporal analysis of road collisions involving vulnerable road users (VRU), using a crash database of the Turin municipality for the years 2006-2020. In the urban road network, collisions tend to cluster at specific points, also called hotspots or hazardous road locations (HRL).Therefore, it is of fundamental importance to identify them in order to (i) define their dependence to road geometric, environmental and operational factors, and (ii) intervene with safety countermeasures. In this study, Turin’s collision data was obtained from the Italian National Institute of Statistics (ISTAT) database. To identify HRL and then intervene, all VRU collisions and collisions related to VRU sub-categories (i.e., pedestrians, bikers, moped riders, and motorcyclists) were evaluated separately. Then, some descriptive statistics of crash data were evaluated by considering weather, temporal, collision outcome (i.e., injury or fatality), gender and age factors. In a second stage, (i) distance- and (ii) density-based methods were applied to detect clusters in the urban area. In particular, distance-based methods, i.e., the nearest neighbour analysis, and G and F functions, revealed that VRU crash data was clustered. Then, density based methods like the Kernel density estimation revealed the HRL in the network. Analyses were carried out using QGIS and ArcGIS Pro software which are based on Geographical Information Systems technology. In the third and last stage, the analysis was performed on a specific urban corridor to highlight the presence of false positive and false negative HRL taking into account the temporal distribution of crash events. Therefore, data of Corso Vittorio Emanuele II (one of the major corridors in Turin) was selected to this scope. A proprietary Python code was developed on ArcGIS Pro to perform this spatio temporal analysis.