What is big data?
Big data has become one of the fastest growing technology trends of recent years. Put simply, big data is a term to describe massive quantities of data that usually surround human behaviours. In a world which now produces so much data, traditional analytical methods have become insufficient and now require processing through computerised methods to spot trends and patterns.
Many industries are benefiting from big data analytics including IT, Finance and Sports. For example, rugby uses big data to improve the performance of their team by combining real-time and historic data on speed, agility and impact through wearable technology. This data is then used to make informed decisions which can change the result of a game. In health, big data is used to monitor the likelihood of contracting certain illnesses and diseases, and in some cases, how disease and epidemic will spread across geographical locations. The most notable case of big data’s role in health is Google Flu Trends.
The search engine giant collected Google search queries and combined this with location data in an effort to predict when and where the influenza virus would be prevalent.
The automotive industry is no exception to the rapid development of big data. Today, OEMs, fleet owners, and car insurance companies all rely on big data to increase the efficiency of vehicles, lower the risk of accidents, optimise supply chains and more.
Trakm8 collect approximately three billion miles worth of vehicle and driver data every year. The range of data includes everything from GPS positioning, speed and acceleration to battery voltages, tilt angles, braking, throttle and engine revs. This data is then stored in secure servers ready to be analysed by big data scientists in order to identify how specific driver behaviours impact on fuel economy and accident risk. In addition to this, Trakm8 is also developing algorithms to anticipate vehicle service and repair requirements based diagnostic trouble codes (DTCs) and battery health.
The benefit of having such accurate data enables users to take preventative measures to keep unwanted vehicle downtime to a minimum. Fleets receive notifications when vehicles require maintenance, such as charging a battery before it runs flat or cleaning diesel particulate filters.
All this information can be accessed through a one-touch app that works across all mobile-friendly platforms, making it speedy and simple for a user to ensure that they and their vehicles remain road legal at all times.
For insurers, it is important that we provide accurate crash data when our telematics unit detects an impact however with the UK’s roads littered with speed bumps and potholes, we must be able to distinguish whether an impact poses risk to a driver or is just an inconvenient everyday event. Our big data scientists have created intelligent algorithms that can differentiate between an impact from a genuine road traffic accident or an everyday jolt caused by something such as a rumble strip or poor road surface.
Trakm8’s algorithm mathematically classifies events by analysing its signature and comparing it against our database of known signals. We use machine learning techniques such as linear regression to provide a measure of confidence for each confirmed match. The algorithms themselves are a closely guarded secret are continuously being improved to provide unparalleled insights to our fleet and insurance partners.
Share this article: