THP Now Using Predictive Analytic Software to Keep Roads Safe
When it comes to the traffic safety tools at the disposal of the Tennessee Highway Patrol, things like enhanced radar systems, faster squad cars and even increased patrols likely spring to mind. Interestingly enough, however, the THP is now relying on something considerably more sophisticated to keep the roads and highways safe: predictive analytic software.
Over the course of the last six months, the agency has been running a pilot program called “Crash Reduction Analyzing Statistical History” — or C.R.A.S.H. — which breaks down the entire state into squares measuring five-by-six miles and provides traffic risk predictions for every square every four hours.
“So it might show that between 6 and 10 p.m. the probability of a serious crash is 68 percent in this block,” said THP Colonel Tracy Trott. “And that’s where the captain should direct his resources.”
As if this wasn’t remarkable enough, C.R.A.S.H. allows all manner of factors to be analyzed from weather patterns and crash history to special events/festivals and even football games, and provides adjusted car accident projections.
Here, the advantage of C.R.A.S.H. is that it enables the THP to send additional troopers to potential problem areas, such that they can either work to prevent crashes from occurring or be in the vicinity should any problems arise.
All this, of course, begs the question as to just how accurate the $243,000 program funded via federal grants has proven in predicting trouble spots.
THP officials have indicated that C.R.A.S.H. has proven to be accurate as much as 72 percent of the time. Furthermore, while concrete data concerning the impact C.R.A.S.H. has had on reducing traffic fatalities remains unavailable, the THP has indicated that the overall number of traffic deaths is down 5.5 percent from the same time last year.
“You can’t predict anything 100 percent,” said one of the statistical analysts working with C.R.A.S.H. “You have some days when the predictions are right on, and other days when they’re way off. Mainly what you want to look for is whether the performance of the model is acceptable. And collectively, it’s been very good.”
It should be noted that the THP has also created a C.R.A.S.H. model dedicated solely to identifying spots where there are more likely to be drivers impaired by alcohol and/or drugs present.
It will be interesting to see if the THP makes C.R.A.S.H. a permanent addition to its traffic safety arsenal and, more significantly, how accurate it proves to be over the course of an entire year. Regardless, it’s encouraging to see the agency willing to put such a premium on preventing car crashes.
Source: Government Technology, “Software predicts when, where accidents occur on Tennessee highways,” Shelly Bradbury, Aug. 4, 2014