Tornado Chart with Radically Different Scaled Data

Hi reikerpg 

technically, with standard native Power BI visuals, you are not really going to get the same result cleanly when the two measures are on radically different scales. The issue is not only building the mirrored shape, but doing it in a way that is still readable and trustworthy. In practice, if someone wants a true result like the screenshot, they would usually need a code-based approach such as Deneb, HTML, R, or Python.

Without those tools, the closest practical workaround is something like a table or matrix with conditional formatting / data bars, similar to the example shown above. That approach can imitate the layout and still keep the numbers readable.

 

 

That said, I would also question whether this is the right visual direction in the first place. This kind of chart is problematic not only technically, but also from a UX and data visualization perspective. When two very different measures are placed into one mirrored chart, users naturally read them as if they belong to the same visual scale and can be compared across the center, even though they really cannot. The left side can also be interpreted as negative because it behaves visually like values plotted to the left of zero, which adds another layer of confusion. So even if you manage to mimic the look, the result is still quite hard to decode correctly.

If the goal is to keep the tornado-like layout, I think the safer compromise is exactly the kind of solution shown here: a table-style layout with two bars, but both read in the natural direction rather than fighting each other around a fake center line. It keeps the pairing by category, but avoids suggesting a false left-versus-right comparison.

I wrote about this issue here as well, in case it helps: https://community.fabric.microsoft.com/t5/Power-BI-Community-Blog/Why-Tornado-is-a-disaster-even-as-a-graph/ba-p/3369539

The pbix with the example is attached