Skip to Main Content
(Press Enter)
man explaining automobile safety-related big data analytics

Use Big Data
to Keep Drivers

Imagine the volume of data that GM collects every day. Across the world, we gather chats through our call centers, conversations with OnStar representatives, logs when people bring their cars into dealerships and information from customer surveys. All of this information translates into more than a billion rows of data. A small fraction of this data — approximately 1 percent — contains critical insights about possible emerging safety issues. If we sort, analyze and process data in the proper way, we can access this information, helping to eliminate safety issues and potentially saving lives.

To accomplish this task, a collaborative GM team recently developed Rapid Algorithm Prototyping for Threat Recognition (RAPTR), a software tool that uses advanced machine learning and natural language processing to take unstructured text data from call center logs, customer feedback, warranty claims and surveys to find and focus on potential safety issues through our Safety and Field Investigation (SFI) process.

For example, RAPTR can take text phrases like “I keep on having issues pairing my iPhone with my Yukon. My phone dies” and “My G6’s seat belts won’t retract, I could die in an accident” and “There’s a weird smell coming from my AC, especially right before my car dies,” and this determines that one is a tech issue, the other a seat belt failure and the other a stall, and recommends the appropriate action to address each.

Developing this innovative tool required the input of GM employees from across Vehicle Safety and IT Global Data AI & Analytics Services. For their efforts, the team won the GM Chief Data and Analytics Office’s 2019 Analytics Excellence Recognition award.

two men discussing automobile safety-related big data analytics