Indiana Launches Predictive Crash Tool for Citizens, First Responders

A new Web-based tool allows drivers and first responders to play a more active and calculated role in avoiding and predicting traffic accidents.

by / November 15, 2016

Indiana has unveiled a new online resource to assist drivers and first responders with predicting and avoiding traffic accidents based on a range of available data: the Crash Prediction Website.

The website, announced late Nov. 14, is a coordinated effort between the Indiana State Police and the Management and Performance Hub (MPH) — an undertaking by state agencies, the Indiana Office of Technology, and the state Office of Management and Budget, which provides centralized data sharing, correlation and analysis for the state in areas where multiple agencies must work together.

The Crash Prediction Website maps the probability of fatal and nonfatal traffic accidents, which have been on the rise since 2011. According to statistics cited by MPH, fatal traffic incidents increased by 7.5 percent between 2014 and 2015.

Maj. Mike White, CIO for the State Police, said its value extends beyond simply mapping available data; it also has the potential to save lives and boost public awareness of roadway safety.

"We can look at what I call flat stats, where you get this data one-sided from one data source," he said, "but if you use multiple data sources and really get smart with it, it [shows] you a picture that might surprise the heck out of you."

The idea for the site came from a similar data analytics undertaking launched by the Tennessee Highway Patrol, White said, adding that they provided valuable background, but didn't quite have the breadth of data available to the Indiana team.

Data from more than 2 million crashes, dating back to 2004, laid the foundation for the forecast model — but not all of those crashes are represented on the map, White explained. For clarity and ease of use, only pertinent crash information is included on the platform. 

The predictive portal shows the probability of accidents across the state within three-hour windows throughout the day. Risk decisions are based on the combination of weather, traffic, road conditions, time of day, historical information and census data.

The likelihood of an accident is then ranked from very low risk to high risk, which is color-coded on a map of the state. Blue indicates a low probability, yellow indicates a moderate probability and red indicates a high probability.

"Even if we get one person to pay attention to how they drive and their environment, and think, 'Hey, I better be more observant when I'm driving through here,' if that saves one life, that's worth it," White said. 

A deeper dive allows users to identify the details of injury and non-injury accidents, which are represented as red and gray dots on the map.

White said the platform's self-learning model alows for updates when new data is entered every three days. While the tool offers a fairly accurate forecast of events, he said he compares it to local weather predictions — and even educated guesses can be off from time to time.

"We listen to forecasters, we adjust what we wear everyday, we bring an umbrella if they say it's going to rain. Are they accurate 100 percent of the time? No. But we pay attention and it causes us to change our behavior," he said. "As a law enforcement officer, I want people to change their behavior. ... If this causes somebody to pay attention ... that's worth it."

As it stands, the crash map is in the input-gathering phase of its life cycle. While White and the rest of the Indiana team are confident the public and state partner agencies will find it useful, he said he hopes to gather input and continue to improve the platform in the months to come.

The CIO said the team has met with other agencies and committees, including the Indiana Department of Transportation, to review the tool and discuss potential applications moving forward.

"We're not going to say how to use this tool. You tell us how you use it," he said. "If they want something like an upgrade later, we're entertaining that. I want this to grow and get better."

Officials hope the crash map will not only better inform the traveling public of roadway risks, but will also help to improve the application of proactive policing and medical resources. A 1 percent reduction of all crashes, according to data from the Indiana Department of Transportation, could result in savings of up to $35 million per year for Indiana travelers — and also protect the lives and well-being of motorists statewide.

Eyragon Eidam Assistant News Editor

Eyragon Eidam is the assistant news editor for Government Technology magazine, and covers legislation, social media and public safety. He can be reached at eeidam@erepublic.com.