Analysis of Data from Insurance Industry Using OLAP and Data Mining

The aim of this thesis is to demonstrate the use of OLAP and data mining, in order to find interesting relationships in real data from the insurance industry. Although these methods are different and with other case applications, their combination can achieve a synergic effect. The thesis is divided into theoretical and practical part. The theoretical part introduces the areas of Business Intelligence (BI) and Knowledge Discovery in Databases (KDD). In this part, the basic principles, definitions, methods of analysis and methodologies are described. Furthermore, the differences between OLAP analysis and data mining are defined, followed by the explanation of a method, in which these two analysis techniques can be combined. The last theoretical chapter is devoted to analytical tools that are used during the analysis, namely Power BI for OLAP analysis and the academic project LISp-Miner for data mining. The practical part contains an introduction to the insurance industry, description of the dataset, its preparation and the process of data analysis. The analyzed dataset belongs to Respect, a corporate insurance broker and contains processed insurance events. At the end of the practical part of the thesis, the obtained results of analysis are evaluated. An interview is conducted with a data analyst from the company Respect. Based on this interview, it will be assessed, whether the findings of the analysis are relevant. In the closing part, the whole thesis and the results of the analysis are summarized.