Hi,
I need to order a set of data based on KPI value (ordered) defining 4 different groups equally distributed (it's not a quartile because).
In attach you'll find a QVW example in which I need to develop a straight table like the results attached in the XLS file.
The main goal is to group dataset into 4 different cluster each on with the same number of customer.
So the procedure must be in theory something like that:
1) Calculate the KPI for each row
2) Order row by KPI (DESC)
3) Start from the first row and count each row with a unique and incremental number
4) Check if the unique number divided by the total number of rows is less than 0,25 (group1), less than 0.5 (group2), less than 0.75 (group3), less than 1 (group4)
It doesn't matter if tow rows has the same KPI.
Rows with the same value of KPI can be associated with two different group.
I have to calculate this directly in the frontend (I have a lot of complex KPI to calculate this in the ETL) and it must be independent by user selection (I mean the group assigned to a specific row must not change).
Do you have any idea?
