Wedding Woes
Options

# I need help with a Venn diagram. :(

member

And you guys are smart.

I am classifying people according to 4 attributes. Each person has at least 2 and as many as 4.  Total # of possible permutations is 10, though the data set only uses 8.

If the attributes are A, B, C, and D, the data set is:

ACD (1)
AC (1)
ABCD (3)
ABD (2)
BC (2)
BD (3)
BCD (1)

I found this venn diagram for a set of 4, but it doesn't really apply, since the outermost regions (representing a single attribute) don't apply at all in this case.

http://en.wikipedia.org/wiki/File:Venn%27s_four_ellipse_construction.svg

Thoughts?

## Re: I need help with a Venn diagram. :(

• Options
edited December 2011
Can you remove one of the circles so you get the eight sections you need?
• Options
member
edited December 2011
There has to be an easier way to do this than a venn diagram. The purpose of a venn diagram is to provide an easy visual, and if it's this difficult to design, then it will be even more difficult to read. Can you make something more like a comparison chart with attributes across the top, people down the side, and filled in squares for each of the attributes they have?
• Options
member
edited December 2011

Yeah. I have the table as a backup.  And I can do totals, too (A total = 11, B=11, C=8, D=14, but it doesn't add up to the total N=17 because some of the categories overlap).

I managed to get one of our visual designers working on it, so maybe they can make some magic happen.  In the meantime, I have teh 4-ellipses as a placeholder. It doesn't look too bad, though, as you say, it's not exactly something they can glance at and immediately discern what is going on.

It wouldn't surprise me at all if we killed it entirely and just went with the table.

Thankfully, it's not due until midday Tuesday, so there's a little bit of time to ruminate.

• Options
edited December 2011
I like the notion of a table, showing all of the different combinations (and the number and the percentage of your sample that each category represents).
This discussion has been closed.