You're describing the type of filtering that QlikView offers. I was asked to emulate this cascading filters behavior and was never able to get it close enough for the client to accept it. One year later the same client came back and said he wanted to start from scratch using Tableau's strength and features. Of course, every iteration I produce that used Tableau filtering was rejected with, "I can't believe Tableau can't do something so simple."
Hope you haven't accidentally run into the same guy.
Hey guys, thanks for the replies! I blinked and two weeks went by...sorry for the belated response.
Shawn, I got a good laugh from your comment because someone says those exact words in almost every Tableau meeting we have. I hope that they implement this in future releases, because it's maddening to not have it available!
Yuriy, your workbook is very helpful. Thank you for sharing that idea as I learned some new things from it. (Haven't had a chance to use Include/Exclude statements yet, so I'm going to have to study them for a bit.)
Hi Yuriy, I wanted to see if I could pick your brain again about getting a feature to work. I'm attaching a copy of my workbook. (Yes, I am still working on this **** thing.)
I made good progress with getting the filter to work, based on your input. In the middle bar graph, it shows the same value for all 3 geographic levels. When the user clicks on the map, filters the 3 bars to the region, county, and municipality that was clicked.
I have this filter action working the way that I want, but I can't seem to figure out how to get combination of the parameter calc and the FIXED LOD calc to return the right values across the board. The list parameter has 20 measures; each one is a field. What I really need to be able to do (I think) is call MIN([Field]) where County = Foo and Municipality = Bar.
SUM() is working (mostly) for the absolute values, but it's returning the wrong numbers for the percentages, and that's kinda where i am stuck.
If you have any ideas or suggestions, I will be eternally grateful.
ACS for forums v5.twbx 1.3 MB
The 20 Measures you've choosen via Parameter
are all hard-coded in the dataset -- including Percentages.
The Percentages are valid when selected at the Row Level,
but aggregating them (in any way), the result is wrong.
To get correct numbers for Ratios, one should calculate them as
SUM( [Numerator] ) / SUM( [Denominator] )
where both Numerator and Denominator are additive values
(mean they can be summed up along any Dimension).
For the Percent Difference the calc could be as above with a little mod:
SUM( [Numerator] ) / SUM( [Denominator] ) - 1
Hope you get the idea.
As an example, I've made changes to your workbook,
making the calculation # 04 (from the Measure Selector)
as described (for both County Graph and Multi-geo Graph views).
Please find the attached.
Please look at these views and the calcs named as [YF : ....].
Hope it could help.
ACS for forums v5_YF.twbx 1.4 MB
Thank you very much for the reply! I'd been thinking...I should probably be calculating this on the fly, but I was asked to hard-code everything first and see if I could get it to work that way. I was stuck in "Excel logic" and trying to figure out how to force it to do the equivalent of a cell lookup. This is so much simpler! (Forest for the trees, etc. *facepalm*)
Thanks again for your help.
Catherine, you're welcome.