Azuga’s Driver Score is a comprehensive assessment of accident risk and is based on contextual driving behavior as measured through Azuga’s OBDII device. GPS data along with accelerometer variations throughout a trip are overlaid with rich multi-dimensional data encompassing real-time weather and data on places to provide context. State-of-the-art advanced data processing and machine learning algorithms have been deployed using Azuga’s Fleet Safety Intelligence platform to develop a Driver Score, which is a normalized metric used to enable unbiased comparisons of driver performance within a group or fleet. This score helps in identifying specific behavior coaching required for drivers as well as to reward safe drivers. Scores can also be used to compare a driver’s performance over time.
The Driver Score consists of component scores comprising Braking, Speeding, Cornering and Acceleration behaviors. Where Azuga’s Fleet Mobile is used, Distracted Driving scores are also computed, and for vehicles that support seat-belt use, a Seat-belt Score is also generated. The scoring algorithm factors in multiple parameters that are known predictors of accidents such as vehicle Class, time-of-day (day versus night driving), day-of-week, magnitude of events, posted speed limits, duration and frequency of events, etc. These spatio-temporal parameters are weighted according to the risk they pose.
Azuga’s Insurance Scores
Azuga also provides an Insurance Risk Score that includes Exposure risk in addition to Behavioral risk. Other parameters like driving during congestion, driving on unfamiliar roads and high frequency of short-trips are aggregated and used to assess total accident risk for each vehicle over the time period of consideration. Risk scores are assigned to zip codes weighted by historic accident rates and real-time weather and traffic conditions.
Scores for each driver are normalized and calibrated using typical driving behavior across all fleets in Azuga’s database. These scores enable carriers to accurately determine an individual driver’s level of risk within weeks of program inception.
Azuga’s Score and Correlation with Accident Risk
Accident models developed with real accident data from several enterprise fleets comprising over 6000 vehicles have shown that Azuga’s Driver Scores are strongly correlated with accidents. The risk of a preventable accident decreases by 26.3% for a 10-point increase in Azuga’s Braking scores. Similarly, the risk is reduced by 21.7% for a 10-point increase in Speeding scores. A 10-point increase in the overall Driver Score, which is a weighted composite of individual component scores, corresponds to a 11.5% decrease in accident risk. If Exposure Scores are considered in the model (for Insurance Scoring), a 10-point increase in the overall Insurance Score decreases risk by 48.6%. More details on these models are available in Azuga’s Accident Risk whitepaper.
In addition to event based scores, Azuga also offers scores using high-frequency data (every second speed and accelerometer data) incorporating Exposure Risk along with Behavioral Risk suited for insurance assessments. The high-frequency data provides a more nuanced evaluation of driving that takes into account less pronounced driving variations.
Overall Driver Score
Individual component scores (braking, speeding, acceleration and idling) are weighted to form a composite weighted average that makes up the ‘Driver Score’. This overall Driver Score provides a good indication of the level of risk associated with a worker’s driving: