Forum Discussion

Rsanjuan's avatar
Rsanjuan
Icon for Advocate III rankAdvocate III
10 years ago
Solved

Calculating Average of Average help

Hi,

 

I have a scenario where I want to compute the total average of one person and another person based on what is selected on a slicer.   Here is an example:

 

1.  Selected one person where the total average PM rating is 4.65

 

2.  Selected another person where the total average PM rating is 4.82

3.  I want to get the average of those two average totals, but it's still taking the number of jobs into account.  I just want to have an average of the average (so 4.65+4.82/2) = 4.735.  Is this possible?

 

  • Rsanjuan's avatar
    Rsanjuan
    10 years ago

    BhaveshPatel and Greg_Deckler,

     

    I was able to figure out the DAX expression for this:

     

    AvgofAvgPM = AverageX(Values(Job[Project Manager]),Job[PM Rating])

     

    Basically taking a unique value in the table project manager, and then getting the average value of each project manager.  Then it takes those values and just averages it.  

     

    Thanks!

     

     

9 Replies

  • Greg_Deckler's avatar
    Greg_Deckler
    Icon for Community Champion rankCommunity Champion

    Is your PM Rating a Measure or are you selecting a PM Average column and use the quick calc to get the average displayed?

     

    If it is a column, you could create a Measure that is effectively:

     

    Average = AVERAGE([PM Rating])
    • Rsanjuan's avatar
      Rsanjuan
      Icon for Advocate III rankAdvocate III

      Hi Greg_Deckler,

       

      It's actually a measure, where Job is the table.

       

      PM Rating = Average(Job[PM Score])

       

       

      • Greg_Deckler's avatar
        Greg_Deckler
        Icon for Community Champion rankCommunity Champion

        How about this Measure?

         

        Average PM Rating = [PM Rating] / COUNTROWS(FILTERS(HoursWorked[Person]))

        Basically, FILTERS returns a table of the values of the filters, we count the number of rows and divide the existing measure by that count.  Not sure if this gets you there but something along these lines perhaps. The problem is that you can't really aggregate Measures but I *think* this will work for your use case.