Edit

Share via

Tutorial: Create a decomposition tree with a Power BI sample

APPLIES TO: Power BI Desktop Power BI service

In this tutorial, you start with a built-in Power BI sample semantic model and create a report with a decomposition tree. A decomposition tree is an interactive visual for exploration and conducting root cause analysis.

A decomposition tree is also an artificial intelligence (AI) visualization. You can ask it to find the next category, or dimension, to drill down into based on certain criteria. For more information, see Create and view decomposition tree visuals in Power BI.

Screenshot shows the completed decomposition tree.

Note

You can also create a decomposition tree in the Power BI service by following similar steps. For simplicity, the rest of this tutorial demonstrates the process in Power BI Desktop.

This tutorial shows you how to:

  • Download the Retail Analysis sample and open it in Power BI Desktop.
  • Create a decomposition tree.
  • Do root cause analysis on your data in the decomposition tree in Edit mode.
  • Save the report and continue root cause analysis in reading view.

If you want to get familiar with the built-in sample in this tutorial and its scenario, see Retail Analysis sample for Power BI: Take a tour before you start.

Prerequisites

Tip

You don't need a Power BI license to explore the samples in Power BI Desktop.

Open the sample in Power BI Desktop

  1. Download the Retail Analysis Sample .pbix file.

  2. Open the Retail Analysis Sample in Power BI Desktop. The report should open in Report view.

    Screenshot shows the report view with the option to choose visualizations.

Create a decomposition tree

Now you can create a decomposition tree to analyze the sample data.

  1. In the Visualizations pane under Build visual, select the Decomposition tree icon.

    Screenshot shows the visualizations pane with the decomposition tree icon highlighted.

  2. Decomposition trees can get wide. Select the focus mode icon or drag the edge so the visualization fills most of the page. In this example, we're in focus mode.

    Screenshot shows the resized decomposition tree.

  3. Expand Sales > This Year Sales and then select Value. Power BI adds the corresponding value to the Analyze box.

    Screenshot shows Value for This Year Sales added to the Analyze box for the decomposition tree.

    Decomposition trees analyze one value by many categories, or dimensions.

  4. Next, select each dimension field that you want added to the Explain by box. Let's select these fields:

    • District > DM (District Manager)
    • Item > Category
    • Item > FamilyNane (Family Name)
    • Item > Segment
    • Store > Chain
    • Store > PostalCode
    • Store > Store Type
    • Store > Territory

    Screenshot shows the fields added to the Explain by box.

    Add as many as you want, in any order. You can use them or not, in any order, in the decomposition tree.

Analyze in the decomposition tree

Now comes the analysis part.

  1. Select the plus sign (+) next to This Year Sales and select High value.

    Screenshot shows High value selected for analysis.

    That means Power BI uses artificial intelligence to analyze all the different categories in the Explain by box, and pick the one to drill into to get the highest value of the measure being analyzed.

  2. Power BI selects Store type.

    Screenshot shows Store Type selected by Power BI.

  3. Select the plus sign (+) next to Same Store and select High value. Keep selecting High value from the top nodes until you have a decomposition tree that looks like this one.

    Screenshot shows all the highest values selected by Power BI.

    We selected Same store > Chain > Territory > DM > Category. Or select other values yourself, and see what you end up with.

  4. You can delete a dimension by selecting the X next to the corresponding heading.

    Screenshot showing how to delete a level from the decomposition tree.

  5. Save the report by selecting Share > Share or Share > Copy link.

    Screenshot shows the decomposition tree with the options available for This Year Sales.

Now anyone who views your report can interact with the decomposition tree, starting from the first This Year Sales and choosing their own path to follow.

Learn about everything else you can do with decomposition trees in Create and view decomposition tree visuals in Power BI.

Related content

Tutorial: Explore a Power BI sample

More questions? Try the Power BI Community