Koliou P., Yannis G., “Advanced Analytics for Traffic Safety: Key Metrics, Predictive Insights, and High-Risk Junction Identification”, Proceedings of the 20th Road Safety on Five Continents Conference, Leeds, UK, 3-5 September 2025.

This study utilizes data collected through the OSeven smartphone application, which continuously records unsafe driving behaviors, including harsh braking (HB) and harsh acceleration (HA). To enrich the dataset, these behavioral metrics were integrated with traffic volume and road characteristics obtained from the Traffic Management Center of the Attica region and Google Maps. Spatial mapping in QGIS was employed to geolocate unsafe events, allowing for precise correlation with specific road segments and intersections. This integration facilitated a comprehensive dataset for analysis, capturing both individual driving behaviors and broader traffic conditions. The analysis identified key predictors of crash frequency, including speed variability, braking behavior, and junction complexity. High fluctuations in speed and frequent harsh braking events were strongly correlated with increased crash risk.