11 Replies Latest reply on Jul 30, 2018 11:39 AM by Brendon Troy

    "Fixing" "average per month" "bias"

    Brendon Troy

      As you can tell from the number of scare quotes in the subject of this post, I'm not even sure how to name this post, let alone whether or not this is a very dumb question. Nevertheless, here goes ...

       

      I have a dataset of actions (chats) that started in December 2015 and continues to today. Here's a chart of the volume of chats by month and year:

      Screen Shot 2018-07-24 at 11.24.15 PM.png

      If I want to see the most active months, that looks like this:

      Screen Shot 2018-07-24 at 11.27.11 PM.png

      My issue with this chart is that it's biased; there are 50% more Decembers through Junes than there have been Julys through Novembers in the dataset, if that makes sense.

       

      What I really want is average number of chats per month (total chats for each month / number of instances of that month in the dataset, I suppose), but I'm stumped as to how to do that. I imagine that some sort of FIXED (LOD) calculation may be involved, but I can't work out how to calculate it.

       

      I hope that this makes sense and, if it does, that there is an answer that I can use that is not so simple that I'll feel like a fool and/or waste others' time answering it. Thank you so much for reading and for answering if you can!

       

      Edit: I'm attaching an example workbook. Thanks again.

       

      Message was edited by: Brendon Troy