Big data and road safety: Open StreetMap and the territorial analysis of accidents

Abstract

Road safety is one of the main challenges for sustainable mobility policies, aiming to reduce the number of accidents and their consequences. Analysing road accidents is essential to identify risk factors and to develop prevention strategies, based on data. The aim of this study is threefold. First, to utilize OpenStreetMap (OSM) data to calculate road accident, mortality, and injury indices by correlating them with the length of road lanes (in meters). Second, to conduct a territorial analysis to identify high-risk areas, thereby supporting road safety planning and third, to enhance national statistical information by estimating accident involvement probabilities, with the ultimate goal of determining real traffic flows (vehicles/km) and actual risk exposure rates. The approach uses an integrating geographic and statistical data using GIS techniques. The researchers implement a spatial join algorithm to overlay information layers derived from OSM and traffic points (PoT).

The analysis includes a new classification of road segments, updated to 2021, and the application of the "Ranker" tool to generate synthetic indicators. Accident data, provided by Istat and other administrative sources, are georeferenced and analysed to highlight territorial variations in risk distribution.

The main innovation of this study lies in the use of Big Data from OSM for statistical purposes, aligning with Trusted Smart Statistics (TSS) initiatives. The integration of geographic and statistical data overcomes the limitations of traditional risk measures based on resident population or vehicle ownership. Furthermore, the introduction of traffic points refines risk indicators, providing a more detailed framework for accident prevention on a territorial scale.