-
1. Re: Difference of change from average change
Shawn Wallwork Feb 2, 2013 8:40 AM (in response to Andrew Marritt)Andrew, did you post the correct version of your workbook?
So on the first sheet the fastest growing group has an 83% change on their starting position over the duration.
Not on the first tab. Here's the % difference from FIRST():
So the fastest growing group has a 26.23% change since the beginning. I do see an 83% change on your sixth tab "Percentage change for each sub group". Is this the tab you want us to work with?
--Shawn
-
2. Re: Difference of change from average change
Andrew Marritt Feb 2, 2013 9:44 AM (in response to Shawn Wallwork)Shawn,
Apologies, I uploaded an old version by accident. For some reason Google Drive hadn't synced between Parallels and OSX.
this should have 3 tabs. The first one is my primary concern but if tab 1 is fixed I should be able to do tab2 using the same calculations.
The first one should look like this:
Thanks for helping
Andrew
Message was edited by: Andrew M to add image
-
Tenure.twbx.zip 31.8 KB
-
-
3. Re: Difference of change from average change
Shawn Wallwork Feb 2, 2013 10:07 AM (in response to Andrew Marritt)Pinging Richard.
-
4. Re: Difference of change from average change
Richard Leeke Feb 2, 2013 11:13 AM (in response to Shawn Wallwork)You need to calculate the overall percentage difference using the same axis on the dual axis map as the breakdown - which means the same fields on shelves. I've added a calculation that does that using the TOTAL() function, partitioned appropriately. That allows you to subtract the overall from the individual.
Hopefully what I've done makes it clear - if not, ask away - I'll have a look later.
-
Tenure RL.twbx.zip 38.0 KB
-
-
5. Re: Difference of change from average change
Andrew Marritt Feb 3, 2013 3:15 AM (in response to Richard Leeke)Thank-you both so much. If you're in London for ECC this year I'll get you a beer (I used to live in Westminster so know some good pubs).
I was trying to do something with Total but was getting lots of 'can't mix types' sorts of errors.