Pattern Recognition and Risk Analysis in U-Turn Crashes

U-turn crashes pose a significant safety concern, particularly in complex traffic environments where the interplay of roadway features, driver behaviors, and environmental conditions increases the risk of severe outcomes. This study examines U-turn crash patterns using 5 years (2017–2021) of crash data from Ohio, applying correspondence cluster analysis to identify latent risk profiles. The analysis revealed six distinct clusters, including high-risk scenarios such as multivehicle crashes in high-speed zones with poor lighting and impaired drivers, alongside lower-severity incidents in low-speed, low-complexity environments. Major contributing factors include four-way and T-intersections, roadway conditions such as single-lane and multilane configurations, posted speed limits, and driver impairment, all of which significantly influence crash outcomes. The findings provide actionable insights for improving road safety, such as targeted intersection redesign, enhanced roadway lighting, and public education campaigns addressing high-risk behaviors. By uncovering critical patterns and risk factors, this study offers valuable contributions to the development of data-driven interventions aimed at reducing U-turn crash risks and enhancing traffic safety.