Understanding Causative Factors in E-Scooter Collisions through Bayesian Network Modeling

Abstract: Shared dockless electric scooters (e-scooters) have rapidly emerged as a popular last-mile transportation solution in urban and campus landscapes. Despite their popularity, concerns about e-scooter safety have impeded widespread adoption, prompting the need for comprehensive research in this domain. Prior studies on e-scooter safety have often been limited by constrained datasets or controlled riding conditions, hindering a holistic understanding of risk factors associated with e-scooter riding. To address this significant research gap, this study undertook an in-depth analysis by collecting extensive real-world traffic crash data and quantifying safety risks linked to pivotal contributing factors. Leveraging the detailed 2020-2021 Fatality Analysis Reporting System (FARS) e-scooter fatal crash data, this research meticulously examined crash patterns at both segments and intersections. The findings present invaluable insights into the key risk factors, illuminating a pathway for e-scooter service providers, municipalities, and campus administrators to develop targeted mitigation strategies. By proactively addressing these identified safety concerns, this study advocates for the integration of e-scooters as a secure and sustainable shared mobility option within urban and campus environments. These findings not only enrich our understanding of e-scooter safety dynamics but also pave the way for safer and more efficient urban transportation ecosystems.