Kind of complicated, but I tried to explain step by step.
25%%75% percentile calc.
Not sure weighted calc is correct or not,, you can modify based on your needs.
Set table calculation appropriately.
Put Index filter
25-75_percentile_SM_9.3.twbx 20.0 KB
Thanks! That was very useful. Now I managed although since the data volume is very big, the computing time is very high, a few minutes.. if you have any other sugggestion to improve it , I'd be very happy to hear it. Thanks again!
Unfortunately, I don't know alternative method at this point.
There is one other way of doing the calcs, namely,
by using a "simpler" PERCENTILE() aggregation --
if the underlying DBMS can do it (not every one can).
Luckily, Tableau Data Engine can do it :-),
so please find the attached workbook
(I took the one by Shin).
ID-ing every row (each time series event) is important
since the NTile (Filter) calc is based upon FIXED LODs
at the granularity of the datasource (every single row).
A caveat: just because of the FIXED,
the whole thing would be slow either.
25-75_percentile_SM_9.3_YF.twbx 71.9 KB