Hi:
To add to Greg's sound advice, if you want to put the numbers on top of the
bars, why bother with the numerical scale? The entire point of a scale is to
provide a reference for comparing different (sets of) values.
\begin{rant}
And when I see things like this:
> dat.bar
> VAR1
> Category 01 17.5
> Category 02 15.2
> Category 03 10.3
> Category 04 8.4
> Category 05 20.3
I get doubly annoyed, because it is yet another attempt to use a bar chart
to plot quantitative values by factor level. As I mentioned in a private
response today, one of the problems with a bar chart is that it forces the
numerical scale to have origin zero, and this is often neither necessary nor
desirable. A simple line plot that connects the quantitative values between
categories is sufficient, and takes *far* less ink to produce. The purpose
of a statistical graphic is to convey information in a simple, clean,
concise fashion - it is not meant to be a rococo art form. If you intend to
write a function to automate a graphic, please think carefully about what is
meant to be conveyed and the *visually* simplest means by which to convey
it.
\end{rant}
The purpose of a bar chart is to visualize a (joint) discrete distribution.
There are better ways to plot quantitative variables by group; in addition
to the line plot mentioned above, the Cleveland dot chart can be very
effective with many groups or multiple grouping factors. With two factors
and a quantitative response, another option is the interaction plot.
If this weren't the third such example/request I've seen today, I probably
wouldn't be so apoplectic...
Dennis
On Fri, Sep 10, 2010 at 10:44 AM, Greg Snow <[email protected]> wrote:
> See this message and the replies to it (and the replies to the replies,
> etc.):
>
> http://tolstoy.newcastle.edu.au/R/e2/help/07/08/22858.html
>
> In there is a discussion of why you don't really want to do that along with
> better alternatives and examples of the improved plots.
>
> --
> Gregory (Greg) L. Snow Ph.D.
> Statistical Data Center
> Intermountain Healthcare
> [email protected]
> 801.408.8111
>
>
> > -----Original Message-----
> > From: [email protected] [mailto:r-help-boun...@r-
> > project.org] On Behalf Of Antonio Olinto
> > Sent: Friday, September 10, 2010 8:59 AM
> > To: R-help
> > Subject: [R] adding labels above bars in a barplot
> >
> > Hello,
> >
> > I want to make a general routine to draw barplots with numbers plotted
> > above each bar. See the example below.
> >
> > I could not place the numbers on the middle of each bar because I
> > could not calculate the right position of each x-axis tick. axTicks(1)
> > indicated a unitary step, but it does not seem work.
> >
> > I appreciate any help or suggestions.
> >
> > Best regards,
> >
> > Antonio Olinto
> >
> > ======================
> >
> > CAT VAR1 VAR2
> > Category 01 17.5 9
> > Category 02 15.2 20
> > Category 03 10.3 500
> > Category 04 8.4 150
> > Category 05 20.3 5000
> >
> > # Coping data from a spreadsheet
> >
> > dat.data <- read.delim("clipboard",header=T)
> >
> > summary(dat.data)
> > CAT VAR1 VAR2
> > Category 01:1 Min. : 8.40 Min. : 9
> > Category 02:1 1st Qu.:10.30 1st Qu.: 20
> > Category 03:1 Median :15.20 Median : 150
> > Category 04:1 Mean :14.34 Mean :1136
> > Category 05:1 3rd Qu.:17.50 3rd Qu.: 500
> > Max. :20.30 Max. :5000
> >
> > dat.bar <- data.frame(dat.data[,c(2)])
> > row.names(dat.bar)<-dat.data[,1]
> > names(dat.bar)<-c("VAR1")
> > dat.bar
> > VAR1
> > Category 01 17.5
> > Category 02 15.2
> > Category 03 10.3
> > Category 04 8.4
> > Category 05 20.3
> >
> > par(mar=c(12,6,3,2),cex.axis=1.2,cex.lab=1.4)
> > barplot(t(as.matrix(dat.bar)),ylim=c(0,max(dat.data[,2]*1.1)),las=2,yla
> > b="Y
> > label text",col="orange")
> > box()
> >
> > up <- max(dat.data$VAR1)*0.1
> >
> > for (i in c(0:nrow(dat.data))) {
> > legend(0.25+i,dat.bar[1+i,1]+up,dat.data[i+1,3],col="blue",bty="n")
> > }
> >
> >
> > ----------------------------------------------------------------
> > Webmail - iBCMG Internet
> > http://www.ibcmg.com.br
> >
> > ______________________________________________
> > [email protected] mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-help
> > PLEASE do read the posting guide http://www.R-project.org/posting-
> > guide.html
> > and provide commented, minimal, self-contained, reproducible code.
>
> ______________________________________________
> [email protected] mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
[[alternative HTML version deleted]]
______________________________________________
[email protected] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
