Not sure if this becomes a suggestion for a future card capability, or a request in terms of is this possible today*?, but let me put this out for consideration...
I have a book of business... thousands of transactions, each transaction has hundreds of data points. But ten of those data points are key, and i want to know the distribution of the overall book of business across each of those data points. I could build ten cards, one for each attribute, and show the relative distribution within that attribute.
Example: attribute=color; values: red, blue, green, yellow; results: red=35%, blue=30%, green=25%, yellow=10%.
What I'd really like to do is have something like a table card or a Sumo card type of layout, where i can add multiple attributes as series, and stack them on top of each other. Each attribute is a series and represents 100% of the total population, but the view is stratified such that i have a stack of attributes, and each one is split by the values of that category. That's the vertical view...
On the horizontal, i could do different things... perhaps multiple different KPIs (unit count, total value, avg cost, etc.) Or perhaps a time-series so I can see how the mix within an attribute changes over time.
The attached screen shot shows an example with a stack of four attributes, each reflecting the potential subordinate values within that attribute and the relative distribution, across a 6-month time series. This example happens to be at a customer level, so several of the categories show 100% in a single row rather than having greater definition as it would across a larger selection.
*I realize I could probably do this with a series of Dataflows, heavily scripted MySQL, and then bringing them together in a final dataset in a particular order, but that lacks efficiency when it needs to be repeatable in different configurations. Being able to build it as a card with multiple value and series drag & drop capability would be a great feature.