Variance analysis solution for financial controlling in Škoda Auto a.s

The objective of this master thesis is to propose a solution optimizing variance analysis for financial controlling in Škoda Auto by developing a new, uniform, and flexible methodology for variance analysis that would be transferrable in between the controlling departments, and to propose an analytical Business Intelligence reporting solution for visualizing the variance analysis results, calculated by the new methodology, and for facilitating their interpretation. The thesis achieved all the set objectives and emphasized the potential complexity of custom variance analysis solutions. As part of the thesis, a unified methodology for the Škoda Auto variance analysis was created, including a more detailed breakdown of the total variance by containing the effects of dimensions (country, plant, etc.). This methodology was approved by the leaders of the relevant controlling teams and is gradually being implemented, starting with the calculation of variances for a management overview report. In the final part of the thesis, an interactive BI analytical solution in Power BI was designed. It includes the application of the methodology's calculations in the DAX language and enables easier tracing of the origin of variances and correlations between the included dimensions. The thesis describes the development of the BI solution, from the initial study, through analysis and solution design, to its implementation in a test environment using Python, Power BI, and DAX on random and pseudonymized Škoda Auto data. In addition to offering a different and more detailed perspective on variance analysis than usual, this master thesis and the new methodology also benefits Škoda Auto by reflecting their current environment in the solution, unifying the variance analysis methodology in the company, and eliminating doubts regarding calculations behind the resulting variances. The company now has centralized calculations for variance analysis in a methodology that includes all meaningful dimensions and is thus complete in terms of the currently available data in Škoda Auto, but at the same time, thanks to the sequential model of dimensions, it can incorporate additional dimensions or car models in the future. The design of the Business Intelligence solution provides Škoda Auto with a partial plan for developing an effective and accurate tool for visualizing and interpreting the new variance analysis.