How to plot box and whisker?

Hi wsspglobal

According to your description, it seems that you want to understand the logic of above chart, right?

For whisker, you could find that it group on category, then calculate  the max of sampling

For plot, you could find that it also  group on “Category”, and calculate max, min based on “sampling”

So you could modify field in “Category ” and “sample”  based on your requirement.

Best Regards,

Zoe Zhi

If this post helps, then please consider Accept it as the solution to help the other members find it more quickly.

Hi Anonymous - I just came across this post while browsing - I'm the author of the violin plot visual and may be able to help clarify its usage, if you're still experiencing some challenges.

This is documented here, but I'll walk through using your data.

You are correct that the visual should create all statistical measures, based on the data you feed into it. In its simplest terms, you need to be able to provide data at its lowest possible grain for this to be worked out correctly. If any of this data is pre-aggregated then it will not work.

From your example, you are plotting Session as a Category and Amount as the Measure Data. If we only pass these in the the visual, then Power BI will aggregate Amount by Session - as you've already indicated in your screenshot - but to illustrate how this looks to the visual:

You will notice that the Amount is aggregated by the Session. In order for the visual to know more about the distribution of your data, we need to give it every row we can and this is where the Sampling field comes in.

You need to create a unique value for this field so that the visual does not aggregate the measure at a higher level. We'll look at your data and try and find a suitable candidate field.

An easy way to check this is by adding a measure that provides a count for the one you want to check. Let's have a look at Subset:

Note that there is a 2 in each column. This indicates that Subset is too high-level to provide all values of Amount to the visual, and we can see the resulting violin plot:

The tooltip says there's 5 samples, because there are 5 distinct values for Subset, and we actually need 10 because that;s what's in your raw data.

An ideal way to do this is to use something like a primary key in your table to do the sampling on. The puts the data at its lowest level of grain. You can do this in Power Query, e.g.:

1. Add Column
2. Index Column
3. From 1

This creates a unique value for every row in your table, e.g.:

Close & Apply the query and this will get added to your table. Let's repeat the above 'Count Test' using Index:

We can now see that every row has a 1 in it, which means that for every row, the Amount is going to be as per the table and won't pre-aggregate, e.g.:

Now, we should add Index to the violin plot so replicate this behaviour, e.g.:

If we check the tooltip for A, we get the stats calculated, e.g.:

This technique will also work for the Box & Whisker Chart, e.g.:

Hopefully this helps clarify the usage and behaviour of both of these visuals, but feel free to ask anything else and I'll do my best to answer.

Good luck!

Daniel