Stefano Raccagni, Roberto Ventura, Benedetto Barabino
• Proposing a refined framework for road risk assessment during mega-events. • Implementing a three-phase approach: preparation, modelling, and analysis. • Validating models using out-of-sample testing to ensure generalisability. • Setting an appropriate probability threshold to address data imbalance. • Simulating treatment impacts and generating abacuses for decision support. Mega-events are large-scale international occasions that pressure transportation infrastructure significantly, leading to increased traffic demand and higher crash risk. Therefore, special attention is required in the planning phase to ensure safe road management. Despite some research on their pre- and post-event safety impacts, no studies have proactively assessed crash risk during the planning phase by integrating frequency, severity, and exposure factors. To address this gap, this study proposes a refined framework that combines predictive models and risk assessment techniques to evaluate crash risk in mega-event-affected road networks. Its applicability is demonstrated through the Milano-Cortina 2026 Winter Olympics case study, analysing 3 k + crashes on the interested road. Results highlight key risk factors, including traffic volume, access points, and operating speed. Moreover, the recommended risk mitigation treatments are simulated, and their effects are quantified. Finally, reference curves and tables are provided to support decision makers. By enabling proactive safety interventions, this framework provides a decision-support tool for event planners, policymakers and traffic engineers to prioritise safety interventions efficiently and apply risk mitigation strategies, ensuring safer transport management before mega-events.