2 messages in com.googlegroups.google-chart-apiRe: venn diagram - relative sizes inc...
FromSent OnAttachments
barryhunter10 Dec 2007 14:27 
helly17 Dec 2007 11:38 
Subject:Re: venn diagram - relative sizes inconsistancy?
From:helly (marc@google.com)
Date:12/17/2007 11:38:06 AM
List:com.googlegroups.google-chart-api

Hi Barry,

this is a known issue. One we sadly found out ourselves right after the release. We have a fix for it in development and will push it out with the next release.

regards marcus

On Dec 10, 11:28 pm, barryhunter <Barr@googlemail.com> wrote:

I have a chart here, that I think is correct:

http://chart.apis.google.com/chart?cht=v&chs=300x200&chd=t:100,44,0,2...u1:2871

this is visualising therelativesize of two properties (u1 and u2), the labels show the absolute values, and the amount of overlap between the two properties.

I calculate one circle to always be 100, and the other the percentage of the other. So in the above case circle B is 44% the size of A, and a has an overlay of 2.2% - all fine. (C is zeroed out)

However in the cases where B is bigger than A,

http://chart.apis.google.com/chart?cht=v&chs=300x200&chd=t:44,100,0,2...u2:6442 (in this case actually just the two above numbers reversed to avoid introducing another example)

The API seems to internally swap A&B circles so that the red circle is always bigger, however it hasn't swapped the legends so therelativesizesdoesn't
match anymore?

So is this a my problem, a misunderstanding ofvenndiagrams, a shortcoming of the API or otherwise?

In the meantime I think I can just sort the data to always be A as the biggest.