Analyse data across multiple dimensions with the Power BI Decomposition tree

With the Decomposition tree visual in Power BI, you can perform intuitive root cause analysis. This visual also works great for ad hoc data exploration by giving a good general overview of data distribution within a model. What’s also great about this visual is that it is completely interactive so, you can easily pinpoint the bulk of volume within your data.

In this blog, I will use the open-source, Adventure works dataset, to determine the drivers behind the Total Revenue. This will include a stepwise explanation of the visual and its fields.

(Step 1) Select the Decomposition tree from the visualisation pane

This one is easy. I’ve just selected the visual from within the Visualisations pane.

(Step 2) Select fields to be used as dependent and independent variables

For the Decomposition tree, you will need to select the field that you want to analyse which is basically the dependent variable. Furthermore, this field needs to be a quantitive measure. For this example, I have chosen to look at Total Revenue. Then once the dependent variable is selected, you also need to select the explain by fields. The explain by fields are the independent variables contributing to the performance of the dependent variable. Here, I have selected the Year, Category Name, Product Key. And at the most granular level, I have the Product Name.

(Step 3) Choose how to split the data

Once you’ve selected the fields for the visual, you will see a little plus next to it. This plus will help to build the decomposition tree by allowing you to choose how to split the data.

These splits appear at the top of the list and are marked with a lightbulb. All AI visuals in Power BI are marked with this lightbulb.

(Step 4) Check the results from the AI splits

According to the AI split, the Total Revenue is highest when the Year is 2016.

The category that brings in the most revenue is Bikes.

(Step 5) Looking at the most granular field

Given all the product names, what product would I expect to be the biggest driver of Total Revenue?

From the above, it is evident that the Mountain Bike- black, with Product Key 362 is bringing in the Most Revenue

In other words, it’s the average across all products in the entire table. The reason why product 362 is at the top is that the difference between what this product drove and the average across all products was bigger than any other difference in the data set.

In Summary

From the above example, you can see how easy and useful this decomposition tree is towards enriching your data analysis experience.

For other interesting blogs, you can also visit our blog site at: https://databear.com/blog/

Official Microsoft information pertaining to this topic:https://docs.microsoft.com/en-us/power-bi/visuals/power-bi-visualization-decomposition-tree

Our training options:  https://databear.com/power-bi-training/