2 Replies Latest reply on May 31, 2017 12:23 AM by David Savory

    Dynamic Grouping based on LoD Calcs

    David Savory

      Hi,

       

      Only started using Tableau a few months ago so bear with me.

       

      I'm analysing a forum and am looking to see how responses to thread creators correlates with their subscription conversion rate.

      The data available is as follows:

      User ID

      Thread ID

      User Group: Admins & Users

      Sub-Forum: General & Help

      Months Subscribed

      Replies 7D: Number of replies within 7 days of thread creation

      Days into Subscription: Number of days into subscription when thread was created

       

      The data is set up so each data is a thread.

       

      Since I'm focusing on the conversion rate from month 1 to month 2 (let's call this CVR1), I need to limit the threads, and have chosen threads within 20 days of subscribing.

      Conversion rate is a user-level metric, not thread-level, so I need to create a calculation for the average replies within 7 days per thread per user (let's call this RPT7 / User), and then group them (using calculations, like if RPT7 / User <= 1 then '1', etc.). But the grouping only works if the data is fixed, I can't create an include LoD and have the grouping dynamically update depending on the view.

       

      I'd like to display the data so I have RPT7 / User groups on the x-axis, and the CVR1 on the y-axis, like in the General CVR tab in the attached sheet. Except the RPT7 / User needs to account for the sub-forum and days into subscription filters (and whichever other filters I include), but doesn't right now. What I need in the General CVR tab is the RPT7 / User for threads only within the General sub-forum and only within the first 20 days. How do I create a calculation which shows me RPT7 / User that is grouped and updates based on filters?

       

      Hope that's clear enough. Thanks.