Pattern classification on specifics of public sector investments and budgeting principles

The aim of the study is to find patterns in an exact complete data set containing the annual budget of all municipal subjects in the Czech Republic over the past 20 years. The focus of the analysis is on which resources could enable the development of and investment in municipal assets, especially estate property. The financial and real estate analysis (FAMA) method was chosen, which provides comparable indicators to calculate the debt service and other related features of subject performance on a municipal level. These indicators demonstrate whether municipal subjects follow responsible budgeting principles and/or how they utilize their own resources. Comparing similar studies using mentioned data and methodology there is a gap between context of data in time a relation chosen indicators. The reason of obstacle is to put data in time-series and properly analyze them. This appropriate items of indicators are aggregated so all the connections between them and other items are lost. In enormous amount of data study uses classification tools to unfold hidden patterns how does municipal budgeting develop in time without knowing details about each case in context of debt and assets. Study convert time dimension to static indicator of its dynamics a using pure K-Means classification conclude having 6 different clusters which differ each other in some of indicators. Within broader context of those clusters we propose an overview of municipal budgeting strategies. In big cities dominates financing of investment by debt and the rest of clusters differs usually significantly with small impact of their population size that is one of determinants budget income therefore essential budget part.