Decomposition Tree is a Power BI visual that allows data analysis according to multiple reporting dimensions. You can view it as a sophisticated drill down tool. It performs data aggregation and gives the possibility of drilldown according to the given dimensions, and it also contains elements of artificial intelligence that help in the analysis and indicate the solution.
At beginning, we imported a dataset. These are sales transactions in one dairy store. We add the Decomposition Tree by clicking on its icon in the visuals palette and its placeholder will appear on the desktop. In order to be able to generate a report, we need to enter values in two fields:
Analyze– here you need to drag the measure (implicit or explicit), the data over which the aggregation is performed. It can be e.g. sum of quantities sold, turnover, average price, etc.
Explain By– here we drag the reporting dimensions according to which we want to perform the analysis. For example: sales group, classification, manufacturer, brand name, item name …
We have created the measure Promet (Sales Turnover), which is the sum of sold amounts in Serbian dinars. To begin with, we will drag it to the Analyze field. A bar will appear in the visual showing the total value of sales. To explain this information a little better, in the Explain By field, we drag the dimensions KLASIFIKACIJA (CLASSIFICATION), PROIZVOĐAČ (MANUFACTURER), ROBNA MARKA (BRAND), NAZIV ARTIKLA (ITEM NAME). We perform drill down by clicking on the “+” icon located to the right of the bar that indicates the value of sales in relation to the given criteria and select the desired dimension. We repeat the process of adding dimensions and depth search. At the end of the analysis, the Decomposition Tree will automatically generate a line showing the “optimal path”. In our example, it is the item that sells the most for given criteria.
By clicking on “+” for drilldown you can select the name of the dimension, as well as the High Value or Low Value. They enable the analysis of all available fields and finding the optimal path in search of the highest or lowest values.
The Decomposition Tree can also be combined with other visuals that automatically filter it, giving you a better insight into the situation you want to analyze.