11 questions about Tourmaline Labs™, Big Data and driving behavior analytics answered


This interview with Dr. Lukas Kuhn was done by Insurance CIO Outlook Magazine.

Lukas Kuhn

Dr. Lukas Kuhn is the Co-Founder of Tourmaline Labs™, Inc. He designed a platform that provides effortless monitoring of driving, location and behavior through low-power mobile sensing combined with telematics analytics for mobile applications.
Previously, Lukas served as a Senior System Engineer for Qualcomm. He obtained a Ph.D. in Computer Science from the Technical University of Munich. Since then he has published 35 peer reviewed research articles and he holds 14 US Patents.


1. What were the factors that spurred the conception of Tourmaline Labs™? Could you give us a quick snapshot of the company’s history and its evolution over the years?

Tourmaline Labs™’ mission is to make roads safer, commutes less stressful, and transportation more efficient and cost-effective through mass-scale driving behavior analytics. Better data yields more optimal decisions and policies. Transportation is part of almost every aspect of life – yet we monitor and analyze less than 1% of all drives today. This is due to the lack of solutions that integrate well into our daily lives and work on the mass scale. Our goal is to monitor and analyze every drive in the world. We saw two key enabling factors to mass-scale driving behavior analytics: (1) efficient, reliable and consistent data gathering via smartphones that are already in the hands of hundreds of millions of people, and (2) a new generation of driving analytics capabilities that reveal novel information such as who is driving, the quality of driving, and the influence of the surrounding environment on the driver. We have solved both challenges and are proud to say that our platform provides unparalleled insight into the driver’s identity, insurability, risk profile, fuel efficiency, deficits, and many more characteristics. Today, Tourmaline Labs™ offers three core product lines: (1) an end-to-end mobile workforce solution to improve fleet efficiency, profitability, driver coaching, and regulatory compliance by bringing location tracking and driving behavior analysis to these fleets; (2) a big data platform for advanced driver behavior analytics to evaluate driving data from smartphones and/or other telematics data sources; and (3) a software development kit (SDK) and Data APIs that enable developers to add accurate driving data and behavior analytics to any smartphone application.

2. As a thought leader, what do you think are some of the biggest pain points in today’s Insurance arena, specifically in the Big Data landscape? Could you explain how Tourmaline Labs™ is addressing the pain points?

We see the biggest challenges in the insurance arena to be in data processing and data analysis. Most of our customers have access to data sets, but face challenges in turning raw data into actionable key performance indicators that add value to their decision-making process. These shortcomings result in analytical findings that are often considered as random or that poorly correlate with the underlying insurance risk. The challenge is to master the analysis of data across different levels of abstraction. In the domain of turning driving measurements into risk indicators the challenges are the same. It starts on the level of analog data analysis where data cleaning, noise reduction, and calibration are concerned with turning flawed sensor readings into motion data that captures the underlying characteristics of driving.

The next level is concerned with the extraction of driving events and driving behavior tendencies. On this level the challenge is to interpret the motion data correctly to avoid arbitrary motion to be mistaken for driving events or to miss events that have occurred.

On the highest level of abstraction the challenge is to interpret driving events and driving behavior tendencies in terms of good versus bad driving, understanding what is normal and abnormal, or what contributes to insurance risk. This last step is particular challenging because many external factors have to be considered to interpret the data correctly. For example, the vehicle type, regional driving trends, road geometry, local infrastructure as well as other factors need to be considered in order to produce a fair, comparable, reproducible, and consistent driving behavior assessment. Overcoming these challenges and interpreting the data appropriately across various abstraction levels is a core strength of Tourmaline Labs™. We build on advanced algorithms in the fields of artificial intelligence, signal processing, biometrics, and machine learning to automatically determine the appropriate means in which to interpret the data and to fine-tune those interpretations based on the data itself. This technique enables our customers to turn data into models and utilize those models to extract actionable driving performance indicators without experiencing the pain of manually crafting and fine-tuning the data interpretations themselves.

3. The amount of data generated today is humongous and data analytics has acquired a new dimension and importance for any enterprise. Could you tell us in detail how you [navigate] (are using) this big data analytics [challenge] to help your clients in auto insurance?

At Tourmaline Labs™, we recognized early on that data sets are constantly growing in terms of number of features, and hence data analytics is growing in complexity too. Typically this growth is exponential in terms of features. This implies that every new feature makes the driving behavior analytics task far more complex. Hand-crafted models or rule-based systems are quickly reaching their limitations under the ever-growing complexity. This demands a shift in paradigm away from manual crafted data interpretations towards automatically generated analytics that deal with the constantly growing complexity. Tourmaline Labs’™ data processing and analytics platform builds on advanced techniques from artificial intelligence, signal processing, biometrics, and machine learning to overcome this problem. Our platform automatically learns the appropriate interpretations to extract from the data set itself. Therefore, as the complexity of the data set grows the analytics capabilities grow with it. This makes our customers prepared and capable to grow their data sets to include more and more features.

4. An excerpt from your website says, “Automatically collect rich Driver Behavior analytics data from any smartphone with our patented artificial intelligence technology. A world of better drivers with our Telematics SDK.” Could you tell us more about Telematics SDK?

Tourmaline Labs’™ Telematics SDK is a software development kit that enables any smartphone application to automatically collect and analyze rich driver behavior data. The main benefit this SDK brings to our customers is the ease of integration, the readiness to be deployed immediately, the minimal battery footprint during operation, and the reliable and consistent data output. This enables our customers to gather a unique insight into the identity of a driver, when and where a drive takes place, and how a driver is performing in terms of his/her driving behavior. Our goal is to enable everyone to create a scalable smartphone telematics solution by simply building on top of our SDK.

5. Your ‘Driver Scoring Behavior Analytics’ uses data for driver scoring and is of great help for fleets and can be a good asset for ‘Meaningful Measurement of Risk for Insurance’. Would you like to shed some light on this?

Our behavior-based driver scoring analytics enable fair and consistent driver scoring results. Driving behavior is influenced by many factors. Some of these are internal factors that reflect the active decisions the driver makes throughout a drive. Additionally, there are external factors (e.g. local weather, traffic, infrastructure, culture, etc.) that are taken into account to quantify a driver’s behavior. The biases these external factors can introduce are removed from the analysis. However, the driver’s reaction to external factors is an important component of behavior-based analysis and is incorporated in our scoring.

Behavioral biometrics is crucial to providing accurate driver scoring. For example, a trip with two hard braking events might indicate good, average, or poor driving performance depending on the other factors involved. To illustrate, consider two extremes: 1) a driver on a straight road with few vehicles in a region where keeping a large distance from the car ahead is typical, and 2) a driver on a busy city street with poor road layout surrounded by an aggressive driving culture. In the first case, two hard brakes is much harder to justify than in the second case. It is more likely in the second case that the driver had to brake suddenly to avoid the car that just cut in front or the pedestrian that ran across the street.

Our patented technology takes these factors into account to provide a fair and consistent evaluation of a driver’s tendencies. The statistical performance over the entire drive is monitored in addition to looking for specific driving events. For example, a driver that continually brakes hard at every stop sign and speeds off quickly (though does not cross the thresholds set for driving events) indicates more aggressive and less efficient driving behavior than one who steadily reduces speed but may have one or two braking events in a city. Additionally, our behavior-based analytics factor in regional differences, infrastructure (e.g. mountains v. flat land), vehicle type, weather trends, and additional information to compare driving performance across different regions, environments, etc. The driver is evaluated over time to derive a deep understanding of a driver’s behavior. However, individual trips may be evaluated in the same manner to provide insight into a driver’s current trends or to distinguish between driver and passenger.

6. Tell us more about ‘Automatic Drive Detection SDK’ its features and benefits?

Tourmaline Labs’™  “Automatic Drive Detection SDK” is a slimmed down version of our “Telematics SDK”. This SDK will automatically detect, gather and catalog all drives but does not provide any insights into driving behavior. We recommend this SDK to customers that are solely focused on gathering data on when and where drives have occurred, e.g. for mileage tracking.

7. Certainly you have helped many clients by providing big data analytics and consulting services in the insurance arena. Could you cite one or two case studies where in you can mention the challenge a particular client faced, the service you provided, and the benefits that they experienced by using your services?

We are currently working with an insurance provider that is implementing dynamic pricing based on driver behavior data and a set of hand crafted thresholds and rules. One of the challenges this client was confronted with was the ever-growing diversity of drivers and regional driving characteristics in the data set. The challenge was to implement a data processing and interpretation layer for a constantly growing and changing data set. The key to the solution was to move from handcrafted models to automatically derived models that were directly extracted from the data set itself. While this provided a solution that could evolve with the data set, it also presented a completely different challenge – the challenge to trust the data. Hand crafted models, despite their limitations, are often built on principles that are intuitively understood. Data driven approaches can often seem counterintuitive and are harder to explain without discussing the underlying algorithmic science. This makes it more difficult for decision makers to build up the trust needed to release a product into production. As part of our service we work with our clients not only on the technical issues they are facing, but we also consult and guide them along the way to reach their goals.

8. Even the most competent organizations tend to face market competition. What are the strategies employed by Tourmaline Labs™ to thwart this competition, and what according to you are the company’s key differentiating factors?

At Tourmaline Labs™, we are proud on our technical excellence and our ability to make driving behavior measurable and the data accessible, usable, and valuable to everyone. Early on we understood that collecting, cleaning and calibrating driving related motion data and making it accessible would not be enough to provide value to our customers. The interpretation of the data is an important factor to make driving data usable and valuable to our customers. Our behavior analytics platform takes motion data from driving as input and outputs actionable key performance indicators that are fair, comparable, reproducible, consistent, and actionable. This is a crucial differentiator to many of our competitors and enables our results to be used in the decision-making process of our customers. Practically speaking, our platform interprets data while taking external factors such as vehicle types, infrastructure conditions, road geometry, and regional driving characteristics into account. Building on this platform, Tourmaline Labs™ is proud to provide unparalleled insight into a driver’s identity, insurability, risk profile, fuel efficiency, deficits as well as many more behavioral characteristics to our customers and to support and guide our customers in the process of implementing data driven decision-making in industries ranging from auto insurance to driver, fleet, and workforce management.

9. If you were to give an analogy between your personal traits/hobbies and your thought leadership in the company, what would that be?

Two personal traits that maybe best describe the way in which we embrace thought leadership at Tourmaline Labs™ are being good listeners while being persistent at the same time. It is all in the balance. A key to our success is to have the capacity to suspend our own agenda and deliberately and empathically allow our customers to be heard. Based on their feedback and our mission we than endure in the face of challenges to achieve despite difficulties.

10. Talking about the company’s roadmap, where do you see Tourmaline Labs™ in the near future? Please elaborate on any upcoming projects/plans that would benefit your customers in the long run?

Tourmaline Labs’™ mission going forward is to assist and support our customers in mastering any driving related decision process, may it be in the domain of efficient and reliable data gathering via smartphones or in the enhancement of their driving analytics capabilities to reveal novel information such as who is driving, quality of driving, and the influence of the surrounding environment on the driver. The key to success for any of our customers is to implement an end-to-end solution that fits their data and business needs, and Tourmaline Labs’™ future lies in our ability to assist our customers in this process.      

11. Would you like to highlight any other interesting insight that we may have missed in this questionnaire?

Ground transportation is one of the largest sectors in the world. In the US alone, 200 million people spend an average of 80 minutes on the road everyday to drive an estimated 3 billion miles. Just on a single day driving results in 14,000 accidents, 85 deaths and 6134 injuries. Driving behavior is estimated to account for annual losses of $240 billion due to accidents and $121 billion due to inefficient fuel consumption. As responsible and resourceful citizens, we all have an incentive to make roads safer, commutes less stressful, and transportation more efficient and cost-effective by adopting technologies like the ones provided by Tourmaline Labs™. We make mass-scale driving behavior analytics a reality - following the fundamental belief that better data yields more optimal decision-making and policies.

Share this post