Use of open data to improve automobile insurance premium rating

An analysis was performed on millions of quotes by a number of insurance companies, covering a major part of the automobile insurance market in Spain. As expected, it was observed that the range of prices can be partially explained by the variables directly associated with the risk being insured in each case (vehicle specifications, age, driver experience, etc.), although there are considerable differences in some cases that are not explained by the said variables and that have a strong dependency on the geographical location of the risk. By using context data from open sources associated with the different Spanish provinces (vehicle fleet, climate, socio-economic data, among others), the premium estimation models are complemented and the price differences are better explained. The use of this type of data and models can help insurance companies to adapt their premium rating and identify possible market opportunities.