IBM > Case Studies > Tennessee Highway Patrol: Using predictive analytics to help prevent road accidents and save lives

Tennessee Highway Patrol: Using predictive analytics to help prevent road accidents and save lives

Customer Company Size

Large Corporate

Region

America

Country

United States

Product

IBM® Cognos® Business Intelligence
IBM SPSS® Modeler

Tech Stack

Predictive Analytics
Data Modeling
Geotagging

Implementation Scale

Enterprise-wide Deployment

Technology Category

Analytics & Modeling - Predictive Analytics

Applicable Industries

Security & Public Safety

Applicable Functions

Logistics & Transportation

Use Cases

Predictive Maintenance
Traffic Monitoring

Services

Data Science Services

About The Customer

The Tennessee Highway Patrol (THP) is a law enforcement agency established in 1929 and headquartered in the state capital, Nashville. The THP is responsible for enforcing federal and state traffic laws. A division of the Tennessee Department of Safety, the organization provides assistance to motorists, investigates traffic accidents and plays a role in criminal interdiction. As budgets for law enforcement become ever tighter, the THP was looking for ways to work smarter with their limited resources to achieve better results.

The Challenge

The Tennessee Highway Patrol (THP) was faced with the challenge of improving highway safety and reducing accident rates without increasing staff levels. The THP aimed to anticipate and prevent accidents by identifying road accident hotspots, prosecute drunk drivers, enforce the use of seatbelts more effectively, and respond to incidents with the most appropriate resources. The THP sought to enable a new data-driven approach to traffic safety that would look for meaningful patterns in the past and apply them to current conditions to predict future events.

The Solution

IBM helped the THP build a predictive model for traffic accidents using IBM predictive analytics for crime prevention and prediction. The model was fed with geotagged historical crash and DUI data from the previous three years, historical weather data, and data about special events. The model seeks out correlations between incidents and external factors: location, time of day, day of the week, time of the year, public holidays, weather conditions and proximity to public events. Given new data on all of these external factors, the model can then extrapolate forward to predict future incidents. During a six-month pilot phase, the THP focused on geographic areas with the highest propensity for severe accidents, dividing the state up into six-by-six mile squares and predicting traffic risks for each in four-hour increments.

Operational Impact

The predictive analytics solution from IBM helps officers decide how to deploy their troopers for maximum effect, and enables troopers to decide the best routes to patrol for each shift they work.
Troopers can be deployed to problem spots ahead of time, either to prevent the predicted crashes from occurring or, failing that, to be on the scene of an accident more quickly to render better assistance.

Quantitative Benefit

In year one, traffic fatalities in Tennessee fell to their lowest level since 1963.
The state also saw a six percent reduction in traffic accident casualties.
A 34 percent rise in driving-under-the-influence (DUI) arrests.