Case Study: Tennessee Highway Patrol achieves reduced traffic fatalities and improved enforcement with IBM predictive analytics

Using predictive analytics to help prevent road accidents and save lives

The Tennessee Highway Patrol partnered with IBM to improve highway safety without increasing staff. Facing tighter budgets and the need to better anticipate accidents, THP used IBM predictive analytics—including IBM SPSS Modeler and IBM Cognos BI—to identify when and where crashes and DUIs were most likely so troopers could be deployed more effectively.

IBM built a predictive model fed with historical crash, traffic, weather and event data and delivered dashboards through the state’s TITAN system to guide patrols. In the first full year of use the state recorded its second-lowest traffic fatality year since 1963; comparing Jan 1–Aug 15, 2014 to the same period in 2013, killed/seriously injured incidents fell 6% (4,914→4,629), safety-belt citations rose 46% (46,118→67,440), DUI arrests rose 34% (3,973→5,313), and alcohol-impaired crashes decreased 8.9%—demonstrating measurable public-safety gains from IBM’s solution.

Tennessee Highway Patrol

Colonel Tracy Trott