Hi,
>>>>> "Dan" == Dan Bebber <danbebber at
yahoo.co.uk>
>>>>> on Thu, 21 Jul 2005 12:05:45 +0100 (BST) writes:
Dan> Hello,
Dan> I wrote a wrapper for symbols() that produces a
Dan> bivariate bubble plot, for use when plot(x,y) hides
Dan> multiple occurrences of the same x,y combination (e.g.
Dan> if x,y are integers).
Dan> Circle area ~ counts per bin, and circle size is
Dan> controlled by 'scale'.
I'm not answering your question, but still, I need to ask/tell
this:
Why don't use sunflowerplot() instead?
The excellent researchers who invented sunflower plots in the
late 70s early 1980s knew well about "bubble" alternatives and
much about drawbacks of such bubbles
{mainly the perception laws of areas vs lengths ..}
That's why they came up with the sunflowers as improvement ..
See 'References' in help(sunflowerplot)
Regards,
Martin Maechler, ETH Zurich
Dan> Question: how can I automatically make the smallest
Dan> circle the same size as a standard plot character,
Dan> rather than having to approximate it using 'scale'?
Dan> #Function:
Dan>
bubble.plot<-function(x,y,scale=0.1,xlab=substitute(x),ylab=substitute(y),...){
Dan> z<-table(x,y)
Dan> xx<-rep(as.numeric(rownames(z)),ncol(z))
Dan> yy<-sort(rep(as.numeric(colnames(z)),nrow(z)))
Dan> id<-which(z!=0)
Dan>
symbols(xx[id],yy[id],inches=F,circles=sqrt(z[id])*scale,xlab=xlab,ylab=ylab,...)}
Dan> #Example:
Dan> x<-rpois(100,3)
Dan> y<-x+rpois(100,2)
Dan> bubble.plot(x,y)
Dan> ___________________________________________________________
Dan> How much free photo storage do you get? Store your holiday
Dan> ______________________________________________
Dan> R-help at stat.math.ethz.ch mailing list
Dan> https://stat.ethz.ch/mailman/listinfo/r-help
Dan> PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html
Dan> !DSPAM:42df82ae77251002021314!