I really like how Causal allows to swap the category aggregation mechanism.
However, I find one type of aggregation really missing: That’s sort of a weighted average - DIRECTLY in the variable itself, as opposed to using a second variable to calculate the weighted average.
Example:
we model the number of shops we get our products into by % of shops of a given supermarket chain.
We therefore create a variable “% of stores” which has 2 levels of categories assigned: 1.) product, 2.) supermarket chain.
We have another variable hat holds the total number of shops for each chain.
So if we plan to have Product A in 15% of stores of chain X and 30% of stores of chain Y, it would be fantastic to directly see in the aggregation of our % of stores variable, what percentage we actually achieve overall across all chains.
For that we’d have to calculate (15% * total_stores(X) + 30% * total_stores(Y)) / total_stores
That, however doesn’t seem to be possible to put as a aggregation-level formula in a variable that serves as direct value-entry.
My preferred solution here would be to be able to define the formula for the aggregate level even if i’m entering values at the detail level directly.
Or am I overlooking something how I can do this straight in my “% of stores” variable already ?
I can’t access the model you linked - only the cards view which doesn’t show me all the details. Can you exten the sharing rights please ? (or send me a full-edit via PM / support chat) ?
That’s exactly right - however, since there are a ton of such variables, it’s a pity that the weighted averages aren’t displayed “in place” on the aggregation level, but you have to find yet another variable to keep an eye on that.
Hey @FabianK - I think it would be hard to display it in place all in one variable as the weighting amounts for each category item will always be a separate variable. I’ll have a chat with our product team though and see if we can come up with something
