Just realized my spacing on the 'example data' didn't translate
Cleaned up example:
FIELDNAME VALUE1 VALUE2 VALUE3 VALUE4 VALUEMAX
BOB 1 1.4 1.7 3 4
SALLY 2 2.5 3.4 3.7 4
TOM 0.85 1.1 1.5 2.3 3.4 4
MARY 1.0 1.2 1.5 1.6 3.2 4
Hope that helps clear up the type of data I have to play with.
not really - (btw: thanks for cleaning the extra data,). I've played w/ shapes as well - but it makes it hard to view after I add in 10+ rows of data.
I was almost hoping for something like a filled in area chart, or something like that of a bar graph that can have multiple sized bars (almost like the gantt chart where you can have 2 bars on the same row on top of each other.
One of the problems is the SCALE or x-axis needs to stay static and only go to the MAX number - however when I try to stack or overlay; it sums the figures together vs just keeping the numbers discrete (which setting didn't help either).
I found this forum post (http://www.tableausoftware.com/support/forum/topic/calculating-difference-likert-scale-multiple-year) which actually has a lot of the things I need (difference calculations etc..) but I couldn't get it to work with the data set I have; I tried to trace it out but got a bit mixed up.
The sample set I gave might have been a bit misleading - to add a bit; its almost like a survey where the first column is a string listing of categories/questions the rest of the columns are results/figures, with each column being for 1 specific responder. I hadn't had my coffee yet when I posted my first post ;-)
So, to keep with the survey analogy) what I wanted to do is a visual that plots each responder had, on a single row, for a single question. But I also want to do a comparative analysis of one group to the next (which is why I was thinking of an area chart). The 'other' data I have would be a differential analysis (like in the link I posted above) to highlight how one group different from the next.
I am not sure if there is any way of overlaying more then 2 bars (such as in the link above) - if so, being able to do so w/ 4 bars in the same fashion would match exactly what I am looking for - if not than I think I can make the style with the link above work, if I could figure it out ;-)
I am not sure I understand the description of what you are looking for. can you please mock up something that represents the result that you want from the sample data you provided?
Well you can certainly do the stacked bars with multiple series - just put series (in this case Measure Names) on both Colour and Size, and then go to Analysis->Stack Marks and set that to Off. See attached.
Like Joe I don't quite understand what you're after with the rest of the question. As well as a mock-up, maybe provide some sample data which is closer to the structure of your real data - it really helps if there aren't too many layers of translation going on. What I took from your last post was "Bob", "Sally", "Tom" and "Mary" should really be "Question1", "Question2"... and "Value1", "Value2" should be "Bob", "Sally"...
Also, how do you know which responders to compare with each other for the differential analysis?
josh2.twbx 9.0 KB
I apologize - I know my notes aren't that solid on what I was looking for (hindsight is 20/20 and all that). Actually, Richard what you have posted is exactly what I was looking for! I think I can figure the rest out. Thanks to all for the posts and insight - the stack marks option seems to be the missing link for what I was trying to figure out.
I will post a reply to thins shortly with an example of what I was trying to get together - just to close the loop on my confused rambling ;-)