🥫 Food has a nutrition label. Training data has nothing.

Where did it come from? When was it collected? Who verified it?

As AI enters finance & climate, 'data nutrition labels' stop being a nice idea and become a strict requirement.

#DataLabels #AIProvenance

Looker Studio adds background colors for bar and column chart labels #LookerStudio #DataVisualization #ChartDesign #Analytics #DataLabels

Looker Studio adds background colors for bar and column chart labels #LookerStudio #DataVisualization #ChartDesign #Analytics #DataLabels

In case a single value is expected the total would be perfect. That would be wonderful to define an axis max that creates enough space for all datalabels to be rendered correctly.

Yes to “Let the creator decide”.

If I want smaller font than 8, let me.
If I want datalabels to overlap, let me.
If I want padding to exceed your arbitrary maximums, let me.
If I want to disable High Density Sampling, let me.
You guys can greatly improve the app by removing limits, at low cost.

Here's a pic of the two "states" side by side. The measure actually plotted doesn't change...it's an absolute count. But the datalabels change between absolute values and % depending on what the user selects.

(Also, the % values change from 0dp to 1dp when the percentages drop below 1%.)

Whoops, forgot that histograms have no gap between bars.

Experimenting with showing male/female avg on the histogram too.
* Not sure if it works
* Because there's no Bar/Line combo I had to use Reference Lines. There s a lack of formatting options cw "proper" datalabels e.g. font size, background😠

8pt for datalabels. If #PowerBI let ME choose whether to suppress datalabels that overlap even a teeny bit rather than inexplicably suppressing them, then I would use 9pt or even 10. Instead I am forced to make dashboards less accessible in order to make them more accessible and it fucking rankles.

For clarity, I would like to be able to populate these two DataLabels using Visual Calcs.

The *series* that they are labels for *are* generated by Visual Calcs, but I've had to use measures to generate the DataLabels themselves, because the DataLabels Fields Picker doesn't show Visual Calcs in it.

My first use case is to draw in "Average" lines. I did notice that there is a bug when applying measure-driven datalabels to them: If I don't have that redundant "" in the FALSE branch of my DL code it injects values from a field not even in my visual along my average lines. Super weird.

Ah dang it I don't think you can use Visual Calculations to populate DataLabels, because unlike say Chart Titles, Data Labels use a Field picker that looks for things in the model.

@jaypowerbi.bsky.social is this the case? If so, is this purely because of the type of picker that is used for DL?

I think the issue is simply that the Datalabels algorithm isn’t aware of the “Show items with no data” option. Probably a simple fix, @myersmiguel.bsky.social. Probably not a high-priority fix, mind.

Basically the crux of it is that if you want to have separate datalabels for null values as well as zero vaues, you need two series - the one you plot needs to turn Nulls to Zero, the one you use for datalabels should turn nulls to "Missing".

This is too hard, @myersmiguel.bsky.social

#PowerBI