Creating Decomposition Trees in Microsoft PowerBI

If you’ve ever faced 10,000 rows of data and tried to make sense of it, this might just be the article you’re looking for. Meet the Decomposition Tree. It’s a Microsoft PowerBI feature that helps you break down your data into something meaningful, without the need to ugly cry in the corner!

What is a Decomposition Tree?

A decomposition tree is a bit like one of those Russian nesting dolls, but for data nerds. It shows you the breakdown of your numbers by peeling back layer after layer, just like an onion (or parfait, coz everybody likes parfait. Sorry, I couldn’t help channelling donkey. Shut up Shrek!

Anyway, as an example, you might want to break down sales by year, quarter, month and so on. If that’s the case, it’s time for a Decomposition tree. The decomposition tree is standard visual in Power BI desktop available regardless of which Power BI licence you have. You can find it along with the other visuals in your visualisations pane.

When to Use a Decomposition Tree

The decomposition tree is fantastic when you’re actively digging into your data, much like how Sherlock Holmes would investigate a crime scene – but with fewer deerstalker hats and more spreadsheets. It’s ideal for those moments when you scratch your head and think “Why do my figures look like that?”

On the other hand, if you’re looking for a static visual for your reports, this might not be your best bet. Since it’s best for exploring data, it’s not necessarily the best option for reporting. Think data discovery and insights, rather than presentation of said data.

How to Create a Decomposition Tree in Power BI

Ok, now that we have the preamble out of the way, let’s get started.

1. Open PowerBI, then grab your dataset. For my example, I’m using a product sales dataset for a global company.
2. In the Visualisations pane, select the Decomposition Tree visual. It looks like a family tree and a flowchart got together and had a baby. When you first add it, you’ll see an empty visual with a lightbulb icon. (Yep, the lightbulb means it’s intelligent, maybe even smarter than some of your mates, but let’s keep that between us.)
3. Select the metric you want to analyse. In my case I’m using Revenue from my Product Sales
   
   
   
   
4. Now, add your dimensions to the Explain by field. I’ve added Date and Product Category, but I could just as easily break the data down by Order Type or Store Location.
   
   
   
   
5. Now the fun part. Click the little + sign beside the analyse field in your visual. In my example, this is Revenue.
6. Choose the dimension you want to use; I’ll go with Product Category to show a breakdown of the revenue by category.
   
   
   
   
7. Repeating this step by clicking on the + beside Mobile Phones displays revenue broken down by year.

Conclusion

Decomposition Trees are a brilliant feature of Microsoft PowerBI, especially when you want to quickly drill down to explore the details of your data. They may look a little intimidating at first but stick with it and you’ll see just how powerful and insightful they can be.

Authors Note: If anyone asks why you’re spending so much time with decomposition trees, just tell them you’re “conducting advanced data analytics.” It sounds much better than “I’m playing with fancy flowcharts!”