Data-driven analysis of hazmat road transportation risks in Turkey

Abstract

Hazardous material transportation is a complex system involving numerous actors, including staff, vehicles, tasks, companies, legislators, and regulatory mechanisms. The management of various hazards inherent in the system is crucial to prevent accidents that can cause catastrophic damage to drivers, residents, other vehicles, passengers, and the environment. Road transportation is the most commonly preferred mode for transporting hazardous materials domestically. Enhancing the safety of hazardous material transportation operations by road is a primary concern for the sector. To prevent hazardous materials (hazmat) road transportation accidents, it is essential to gather information from accident reports, including accident precursors and outcomes, to improve safety. However, research investigating accidents related to the transportation of hazardous materials on highways is limited. This study aims to determine sequential rules based on temporal dependencies between accident precursors and outcomes, unlike previous studies, using hazmat road transportation accident records within the scope of inland transportation in Turkey from 2015 to 2020. The TRuleGrowth algorithm, a significant task in data mining, was used to analyze accidents and present 75 sequential rules for the most frequently occurring combinations of outcome variables, including injury, fatality, injury and fatality, and damage level. The study’s results are expected to contribute to enhancing the safety of hazmat road transportation and to the literature.