Machine learning model based on detecting and scoring dependencies of different factors of user’s dataset. Points out risk area for better decision making for an insurance industry.
Services provided by Celadon: machine learning, data analytics, Big Data management frontend development, backend development, QA, IT consultancy.
Powered by TensorFlow award winning concept turned into a fascinating MVP. A unique driving hazard assessment analytics. Currently integrated into a several running business and proved its effectiveness.
Millions of people all over the world use the services provided by insurance companies. You can insure practically anything: life, wellness, property, business deals, etc. So it’s no surprise insurance industry is blossoming.
We were approached by a company that had the following idea: to grant the insurance companies more data about their clients in order to minimize any risks. However, the company wanted to cover only one segment: the insurance of cars and the well-being of drivers and their passengers.
The client provided us with a big data set that contained data about the motion of the users’ devices, positioning of the devices during the motion, smoothness of motion, speed, and similar data.
In order to succesfully resolve the issue, the Celadon team faced the following question:
We would never get down to such a project if we were not 100% sure in own skills and knowledge in the field of Artificial Intelligence and Machine Learning. As well, what motivated us was the fact that we got access to an amazingly precise data on the device positioning and motion so we immediately thought of analyzing the manner of driving.
An insurance company wants to get a forecast on the probability of an insured event occurrence with a client. We had a lot of data on the movement at our disposal and used Python and TensorFlow, which is an open library for machine learning. Our developers performed the task brilliantly and offered the client to develop an interface for displaying the analysis results.
In order to clearly visualize the analysis results and obtained forecasts, we used React as one of the most popular and useful libraries for user interface development. It fitted perfectly for the analytical data output.
This data is analyzed in the server and ML technology then builds accurate forecasts of the insured events. The insurance company receives the data coming to the web-interface and makes data-powered decisions that help minimize the risks of financial losses.