Machine Learning Case Study
Octo Telematics, a leader in telematics for insurance companies, is introducing innovations for insurance by aggregating 186 billion miles of driving data from connected cars and using Cloudera Enterprise to predict and model driver risk. Utilizing machine learning and analytics on Cloudera Enterprise, the Octo Telematics platform can now deliver actionable insights and risk modeling to over 100 insurance companies worldwide that enables them to transform insurance including pricing, First Notice of Loss (FNOL), and claims insights.
Octo Telematics’ platform, built on Cloudera Enterprise, aggregates over 11 billion new data points from 5.4 million connected cars and sensors daily, creating one of the largest connected-car subscribed installed bases across the globe. Every relevant type of data—contextual data, driving data, behavioral data, and crash data—is utilized to forecast individual driving habits, improve crash notifications and response, evaluate crash dynamics, and detect fraud.
At Octo Telematics, we are using the power of IoT data analytics to transform the auto insurance industry and essentially be there for people at the moment they need help. With Cloudera, we use IoT and machine learning to drive intelligence and predict outcomes from billions of miles of driving data, and over 438,000 severe crashes from over five million cars – the largest connected-car subscribed installed base in the world,” said Gianfranco Giannella, chief operating officer at Octo Telematics. “With our new Platform-as-a-Solution service offering, cloud is essential to us, but due to regulatory requirements, we also need services on-premises. Cloudera Enterprise natively supports a hybrid environment, so we have the flexibility to run our data where we need to at the lowest total cost. We have a platform for the future.”
Running data between cloud and on-premise environments with Cloudera Enterprise, Octo is able to store, process and analyze data reflecting over 20 million miles of driving per minute using Apache Spark. The data captured includes many kinds of telematics insights, including location, acceleration, braking, idling, collisions and cornering.
Processed data is then made available to various analytics algorithms that revolve around driver risk scoring models, pricing algorithms, accident reconstruction, claims management, crash alerts and analysis, driver behavior, risk modeling and vehicle health – all of which are essential to improving the financial performance of insurers and outcomes for consumers. By building machine learning models on this massive data set, Octo can now formulate more accurate predictions and risk models for individual drivers. Plus, modelers can test new ideas and techniques two times faster and refine on the fly to produce innovative products and services — using data volumes never before possible.
Octo Telematics needed the flexibility, agility, and scale to run its machine learning risk models and predictive applications, both in the cloud and on-premise, to meet service level and economic goals,” said Dave Shuman, IoT and manufacturing industry leader at Cloudera. “Using Cloudera Enterprise to replace a bespoke data platform, Octo has been able to increase its capabilities to scale with the explosion of data volumes and analytic workloads. With Cloudera, they have been able to transform the automotive insurance industry, and these capabilities translate into better outcomes for consumers and underwriters.”
Sign up for the free insideBIGDATA newsletter.
Speak Your Mind