I am running into a problem that looks like it should have an easy solution but I don't seem to be able to figure it out.
I am analyzing a set of time series representing some sort of flow through the system.
Each time series represents a specific sub-flow that is governed by its own rules, but in the end I am interested in the totals (and some sub-totals) of the flows because that is what the system has to deal with.
What I would like to do is forecast each time series individually into the future and then add them up (in various combinations) to get the estimate of future flows.
What the Tableau seems to insist on doing is add up the time series data (as per combination requested) and then forecast the resulting time series.
So, if we call F the forecasting function and my time series x_1 and x_2, I want X = F(x_1) + F(x_2) but I get X = F(x_1 + x_2).
One (clunky) solution is to forecast x_1 and x_2 in tableau, export the result into excel, add forcasted time series in excel to create desired time series and re-import this into tableau to do the analysis and visualization - but this is time consuming and ugly.
Is there a way to get a result I want without leaving Tableau?