Tvorba BI systému pro podporu rozhodování

This thesis is written as a practical project for an educational platform called Hyperskill. Hyperskill is a platform for project-based learning, whose users study programming languages, mathematics, bioinformatics, and other subjects. The company has various content teams. Before the changes introduced in the course of this project, content teams’ producers used to manually calculate and manage employees’ salaries. That took a lot of time, and besides, the absence of automatically generated dashboards prevented the top management from seeing the whole picture and did not let them compare and analyze the situation in different departments. Additionally, despite various people contributing to creation of a content unit – authors of topics or projects, proofreaders, experts, and reviewers – the cost per content unit was not calculated. The only financial information that was available were the salaries of employees. It posed a problem for analysis, since an employee could be responsible for different types of tasks, not just content creation, which was not taken into account. The main goal of this master's thesis was to build a Business Intelligence system that would support management activities and decision-making. Interviews, cross-validation opponents discussion, and job story description were used as methods to determine the problems and the requirements. Based on the established requirements, a new system architecture was designed, multidimensional analysis and modeling were carried out, and a database model was created and described. Modeling and multidimensional analysis were used as methods at this stage. A substantial part of the thesis was dedicated to the implementation of the developed Business Intelligence system. This determines the practical significance of the thesis. First, the author automated the payroll process and converted data to a single, convenient format. Second, she automated the process of budget planning and actual budget collection. Third, she proposed an analytical dashboard design and the metric description for analyzing expenses in different areas, categories, etc, for identifying the cost per content unit, and determining the effectiveness of employees. Experiment, instrument and feature testing, interview, data analysis and training were used as methods at this stage.