R help - Feb 2002 - Smoothed lines over barplots

Greetings.

I'd like to generate a graph that displays two distributions as
side-by-side bar graphs, then plots a smoothed line of the distribution on
top. The idea is to be able to visually compare the distributions.

I've done the following, but the axes don't line up. Any suggestions?

barplot(table(hcd.df$datecat, hcd.df$auth.sum), beside=TRUE,
  legend.text=c('Pre-9/11','Post-9/11'),main='Authoritarian
Sum',
  sub='Chi-Square: 33.88; p<
.01',col=c('yellow','red'))
par(new=TRUE)
plot(table(hcd.df$auth.sum[hcd.df$datecat==1]), type='l',xlim=c(-8,7),
  ylim=c(0,59),col='yellow')
par(new=TRUE)
plot(table(hcd.df$auth.sum[hcd.df$datecat==2]), type='l',
  col='red',xlim=c(-8,7), ylim=c(0,59))


Thanks.

----------------------------------------------------------------------
Andrew J Perrin - andrew_perrin at unc.edu - http://www.unc.edu/~aperrin
 Assistant Professor of Sociology, U of North Carolina, Chapel Hill
      269 Hamilton Hall, CB#3210, Chapel Hill, NC 27599-3210 USA


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

Hi,

On Tue, 19 Feb 2002, Andrew Perrin wrote:

.Greetings.
.
I'd like to generate a graph that displays two distributions as
.side-by-side bar graphs, then plots a smoothed line of the distribution on
.top. The idea is to be able to visually compare the distributions.
.
.I've done the following, but the axes don't line up. Any suggestions?
.
.barplot(table(hcd.df$datecat, hcd.df$auth.sum), beside=TRUE,
.  legend.text=c('Pre-9/11','Post-9/11'),main='Authoritarian
Sum',
.  sub='Chi-Square: 33.88; p<
.01',col=c('yellow','red'))
.par(new=TRUE)
.plot(table(hcd.df$auth.sum[hcd.df$datecat==1]), type='l',xlim=c(-8,7),
.  ylim=c(0,59),col='yellow')
.par(new=TRUE)
.plot(table(hcd.df$auth.sum[hcd.df$datecat==2]), type='l',
.  col='red',xlim=c(-8,7), ylim=c(0,59))
.

The minor problem is that you use par(new=TRUE) on plot() instead of
lines().  lines() automatically keeps the old figure (including definition
of axes).

The major problem is that when ploting as you do: plot(table(..),...) the
plot is assuming the x-coordinates are 1,2,3,.... This is not what is done
by barplot.  Barplot() calculates its own x-values which do not correspond
to the labels.  Luckily you can get its values quite easily.  Simply write

bp <- barplot(..., beside=T)

and bp is a matrix where you have the means of x-coordinates of the bars.
So you can do somthing like

datecat <- rbinom(100,1,0.5)
auth.sum <- rbinom(100,10,0.5)
bp <- barplot(table(datecat, auth.sum),
beside=TRUE,col=c('yellow','red'))
mean.bp <- apply(bp,2,mean)
lines(mean.bp,table(auth.sum,datecat)[,1], col="yellow")
lines(mean.bp,table(auth.sum,datecat)[,2], col="red")

Note that it is safer to use table(...)[,1] instead of table(...[..==1]) --
you get value 0 instead with in the first way but one datapoint less with
the second way.

In fact it was no smoothing here.

Cheers,

Ott Toomet

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._