The November update to Power BI introduced a ‘new’ powerful visual: the Decomposition Tree. It’s an intuitive way to drill down into your data and identify high or low outliers.
In the early 2000s, the Decomposition Tree was a distinctive visual for what was then known as ProClarity: an OLAP tool based on SSAS cubes. After Microsoft acquired ProClarity in 2006, this visual made its way into PerformancePoint. However, with the phase-out of PerformancePoint, this beautiful visual became unavailable. It had, in fact, long been high on the wish list of the Power BI team (and the community).
The Decomposition Tree is now available in Power BI (as a preview feature for the time being).
There are several reasons why the Decomposition Tree is such a powerful visual. For example, the path along which you zoom in (or drill down) remains visible. If you apply a drill-down to a ‘regular’ bar chart in Power BI, it’s harder to keep track of what you’re looking at because the higher-level levels are no longer visible:
When using the Decomposition Tree, the current zoom level remains visible:
In addition, with a bar chart, the zoom path is predetermined (for example: country – state – city): in principle, it is not possible for a user to choose a different zoom path. In the Decomposition Tree, however, this is possible:
In this way, a user can build their own line of reasoning by exploring cross-connections and interweaving various dimensions.
Finally, it’s worth noting that Microsoft has added some additional intelligence to the visual (see Figure 5): By selecting ‘high value’ or ‘low value’ in the zoom feature, Power BI automatically determines which value in the available dimension is the highest or lowest. This makes it easy to identify which dimension makes the biggest difference.
All in all, a really nice addition that was worth the wait!
