R help - Jun 2009 - a plot of stacked boxes

Hello,

I would like to create a plot composed of stacked boxes (squares or
rectangles), where the size of the box would represent the frequency of
observations based on a categorical variable (group), the color would
represent the proportion of success (binary) within that group (outcome) on
a predetermined color scale. Ideally the boxes can be stacked from the
bottom left to the top right based on size. The x and y scales are not
informative. I am experimenting with the rect() function and loops but I
wasn't able to generate an appealing graph.

Thanks

Osman

Here is simulated data set

group<-rep(c('a','b','c','d','e','f','g'),c(100,20,300,40,2,10,40))

outcome<-c(sample(c(0,1),100,rep=T,p=c(0.98,0.02)),

sample(c(0,1),20,rep=T,p=c(0.9,0.1)),

 sample(c(0,1),300,rep=T,p=c(0.99,0.01)),

  sample(c(0,1),40,rep=T,p=c(0.95,0.05)),

   sample(c(0,1),2,rep=T,p=c(0.5,0.5)),

    sample(c(0,1),10,rep=T,p=c(0.9,0.1)),

     sample(c(0,1),40,rep=T,p=c(0.8,0.2)))

df<-data.frame(group=group,outcome=as.numeric(outcome))








Osman O. Al-Radi, MD, MSc, FRCSC
Staff Cardiovascular Surgeon
Co-medical director, Tissue Bank
The Hospital for Sick Children
University of Toronto, Canada

	[[alternative HTML version deleted]]

Hello,

I would like to create a plot composed of stacked boxes (squares or
rectangles), where the size of the box would represent the frequency of
observations based on a categorical variable (group), the color would
represent the proportion of success (binary) within that group (outcome) on
a predetermined color scale. Ideally the boxes can be stacked from the
bottom left to the top right based on size. The x and y scales are not
informative. I am experimenting with the rect() function and loops but I
wasn't able to generate an appealing graph.

Thanks

Osman

Here is simulated data set

group<-rep(c('a','b','c','d','e','f','g'),c(100,20,300,40,2,10,40))

outcome<-c(sample(c(0,1),100,rep=T,p=c(0.98,0.02)),

sample(c(0,1),20,rep=T,p=c(0.9,0.1)),

 sample(c(0,1),300,rep=T,p=c(0.99,0.01)),

  sample(c(0,1),40,rep=T,p=c(0.95,0.05)),

   sample(c(0,1),2,rep=T,p=c(0.5,0.5)),

    sample(c(0,1),10,rep=T,p=c(0.9,0.1)),

     sample(c(0,1),40,rep=T,p=c(0.8,0.2)))

df<-data.frame(group=group,outcome=as.numeric(outcome))








Osman O. Al-Radi, MD, MSc, FRCSC
Staff Cardiovascular Surgeon
Co-medical director, Tissue Bank
The Hospital for Sick Children
University of Toronto, Canada

	[[alternative HTML version deleted]]

I think you are looking for mosaic()

table(df)
library(vcd)
mosaic(outcome ~ group, data=df)

Rich