Data analysis for identifying high chance scenarios of hazardous material highway transportation incidents

Although the transportation of hazardous materials in the U.S. follows rigorous safety regulations, incidents happen during any transportation phases. During past ten years, over 87% of the incidents occurred on the highway. These hazardous material highway incidents (HMHIs) have caused near 760 million dollars of damage, 194 hospitalized injuries, and 110 fatalities. 81% of the fatalities and 87% of the hospitalized injuries are transportation workers. We study HMHIs that caused hospitalized injuries (type-H HMHIs), and those caused fatalities (type-F HMHIs), all during 2008-2017, to develop a thorough understanding of these severe incidents. A distribution-based comparison method is developed for assessing the heterogeneity of any incident sample. Pareto-type distribution charts are generated for each sample of study and its population on various data fields. Then, a series of metrics are developed and used to measure the heterogeneity of the sample through a series of pairwise comparisons of the distribution charts. Results show that both type-F and type-H HMHIs are non-homogenous samples of their population. Moreover, unique features of each sample are identified. Following that, the study further develops a sequentially unfolding strategy for efficiently identifying high chance scenarios of HMHIs. Six scenarios are identified, where assistances provided to transportation workers will effectively lower the chance of type-H HMHIs, type-F HMHIs, or both of them.