- A wealth of information is collected from vehicles’ on-board computer-diagnostic systems. This includes driving behavior, speed, emissions, fluid levels and other operational inputs. By analyzing the interactions of millions of data inputs from a vehicle, our algorithms can identify the patterns and conditions that might have caused a part to fail, and more importantly, tell us why it failed.
- Our predictive-modeling algorithms are able to detect anomalies in the sensors that monitor a vehicle’s on-board tire pressure monitoring system. Vehicles of a specific model year driven in zip codes where the altitude averaged more than 4,500 feet had a higher rate of faulty warning lights, resulting in unnecessary dealer visits that cost both drivers and car dealers time and money. After correcting the sensor software, unnecessary visits declined by more than 60%.