3 Replies Latest reply on May 16, 2017 8:10 AM by Zhouyi Zhang

    Compare average of (parameterized) subset to average of all

    Ken Pattillo

      Hi community,


      First post.  Actually, first viz. 


      If I put a measure (average) on rows and a two member dimension on columns, I would expect two vertical bars, one with the average of the first dimension value (say, male), and the other with the average of the second dimension value (e.g. female).  But what if I want two bars, male and all?  What if I used circles instead of bars, and I want to draw a line between the circles, to imply that the average for males is less than or greater than the average for the total?


      I've been asked to make a dashboard where the user can see some measures (averages) for an entire cohort vs. the average for a subset of the cohort.  The subset is defined by a parameter, minimum GPA.  So, for example, I'd like to see the average GPA for all students vs. the average GPA for those whose GPA >= 2.5.  In the attached data set, I have the HS GPA for 67 students.  I'd like to show a mark for the average of all 67 (2.63), a mark for the average of only those with GPA >= 2.5 (3.022), a label with the correct average on each mark, and a line connecting the two dots.


      It seems to me I need a dimension with two values, subset and all (analogous to male and all).  I can't figure out how to make that from a calculated field.  Or, maybe this is a table calculation, or a LOD calculation, but I don't think so.  Any direction would be appreciated!