With the proliferation of GPS devices, more and more data is available about vehicles in traffic. At the same time, thanks to this growing, pretension appeared as well to use this extracted information for some kind of business purpose.
Data mining is an interdisciplinary field of computer science. With data mining it is possible to extract useful, unknown correlations and new information from huge datasets.
The goal of this thesis is to create driver profiles, using highly accurate and huge datasets that contain GPS information.
CRISP-DM(Cross Industry Standard Process for Data Mining) is a widely known method. It gives well defined frame for knowledge discovery processes from a business point of view. By employing the needed CRISP-DM substeps useful information can be extracted from the raw data.
The thesis follows the steps of this knowledge discovery process. Starting with data understanding, following by data transformation, through modelling, and ending with the explanation of the created profiles.