Lead presenter: Stephen Eick, VisTracks, Inc.
Biography:
Stephen G. Eick is a thought leader in the study of visual analytics and visual business intelligence. As an entrepreneur, Dr. Eick founded VisTracks, SSS Research, and Visual Insights. His leadership and innovation have resulted in numerous best-in-class product awards and a coveted Smithsonian Award for eBizInsights. Dr. Eick also served as Deputy Director, National Center for Data Mining at the University of Illinois in Chicago. He had an extensive career with Bell Labs and managed their Visualization Research Group. Under his direction, 39 patents have been filed. Dr. Eick holds his PhD in Statistics from the University of Minnesota, an MS degree in Mathematics from the University of Wisconsin and a BA in Mathematics from Kalamazoo College.
Visualizing and Predicting Hours of Service Violations
Description
Abstract:
Hours of Service (HOS) regulations are issued by the Federal Motor Carrier Safety Administration (FMCSA) and govern the driving hours of anyone operating a commercial motor vehicle in the USA. The primary purpose of the HOS regulations is to prevent accidents caused by driver fatigue. This is accomplished by limiting the number of driving hours per day and number of driving and working hours per week. To enforce compliance with the HOS regulations, the FMCSA mandated that commercial drivers use Electronic Logging Devices (ELD). ELDs are devices that automatically captures and record a driver’s driving time. ELDs monitor a vehicle’s movement, miles driven, engine hours, and approximate location. In addition, many ELDs, such as those developed by VisTracks®, will monitor the driving hours and warn drivers of impending HOS violations. We performed a study of 293,464 HOS violations taken from approximately 85k drivers and 15k motor carriers. In the study we developed an interactive analytical dashboard, maps of the violations, and animations to help us understand HOS violation patterns. In additional to visualizing violations, we have developed a random forest decision tree model for classifying and predicting when and where violations will occur using historical driving patterns, violation patterns, and weather patterns. The most common HOS violation is when a driver neglects to take a 30 minute break at least once every 8 hours. The time when violations occur most frequently is on Friday afternoons, presumably when drivers are stretching to get home for the weekends. One large motor carrier’s drivers accumulated 10,764 violations over the study period. The most important feature to predict whether a driver will receive a violations is his personal history of violations.