2 Replies Latest reply on Apr 28, 2016 6:43 AM by Petr Safar

    Weighted histogram

    Petr Safar


      I've got an interesting problem.

      Initial point is a histogram based on percentile of dimension Key (calculated in dim. KeyBin) - meaning that count of items in every bin is roughly the same (as is denoted by number of records graph). But the underlying data are biased towards higher values of key (based on value of Weight.)

      The change I need to make is to alter the histogram so that it doesn't contain the same count of records in each bin, but the sum of weights of records falling into each bin should be the same (so then the third graph SUM(Weight) would be the same of all buckets).

      Anybody any ideas?

      Thanks in advance,