1 Reply Latest reply on Jun 27, 2016 8:01 PM by swaroop.gantela

    Deaggregated number of sessions per user.

    Brad Davis

      Hi all,


      I am trying to display the average number of sessions per user per cohort day (defined as the difference between the current date and the first date a user shows up in our system).  I know how to do this in an aggregated way, but not on a user by user basis.  More specifically, the x-axis is going to be the cohort day (day 0 (the users birth date), day 1, day 2, ..., day X), and the y-axis is the number of sessions a user had on that day. For example, if we had 5 users on day 0, and they logged in 2,3,4,4,5 times, then I'd like to display a dot at y=2, y=3, y=5, and two dots for y=4, all wrapped up in a boxplot. Ideally, the two users for x=0, y=4 would be displayed as a slightly larger dot than for x=0, y=(2,3,5).


      I have tried plotting the number sessions for each user by displaying the number of unique session identifiers per user as {FIXED [User Id]: COUNT([UniqueUserSessionIdentifier])}, which is getting me close to what I want, except that it doesn't also convey the differences in the number of users in each of these sub-categories.  Does anyone have any suggestions?