Theme of this thesis is data mining on data of real company Stavebniny DEK ČR. Massive volume of data in intern systems lacks potential without proper usage of data analysis, therefore my goal was to extract interesting information from company’s transactional data.In theoretical part, company Stavebniny DEK ČR is introduced, enlightened the situation on construction material market, followed by description of data mining. Follows description of particular methodology CRISP-DM and its partial ph... zobrazit celý abstraktTheme of this thesis is data mining on data of real company Stavebniny DEK ČR. Massive volume of data in intern systems lacks potential without proper usage of data analysis, therefore my goal was to extract interesting information from company’s transactional data.In theoretical part, company Stavebniny DEK ČR is introduced, enlightened the situation on construction material market, followed by description of data mining. Follows description of particular methodology CRISP-DM and its partial phases: Business understanding, data understanding, data preparation, modeling, evaluation, and deployment.In practical part of thesis there is analysis based on phases of CRISP-DM itself. Five methods of data mining are performed, including 2 unsupervised methods and 3 supervised. Results of analysis are presented, as they would be presented to people, who don’t have experience with data analysis. In the end, a set of recommendation is presented. Results of performed methods led to conclusion, that future analysis should be based on dataset containing more attributes related to customer in purpose of more successful prediction of business cases. Outcome of this thesis could be used by sales department to change parameters of partial business strategy or as foundation for deeper business analysis. |