Paul Johnson
2005-Sep-21 21:42 UTC
[R] controlling usage of digits & scientific notation in R plots; postscript margins
Dear R users:
I assigned students to make some graphs and I'm having trouble answering
some questions that they have. We are all working on R 2.1 on Fedora
Core Linux 4 systems.
1. In the plot, the axis is not labeled by "numbers", but rather
scientific notation like "-2e+08" or such. We realize that means
-200,000,000. We want to beautify the plot. We would rather just print
out the whole, big number. But if we can't have that, we would like
something more beautiful and mathematical, such as
8
-2.0x10
Once the numbers are big, R automagically switches to scientific
notation, and turning the values horizontal does not help on the y axis.
Example:
x <- rnorm(100,mean=100000,sd=100000000)
y <- rnorm(100,mean=10000000000,sd=100000000000)
plot(x,y,axes=F)
axis(2,las=2) #turns y axis numbers perpendicular to y axis
axis(1) # x axis
I realize a person could just divide all numbers by 1 million and then
have a graph with -200 and an annotation (in millions).
On the x axis, we'd even settle to just label the two outermost points
with full numerical values, and then have only ticks between. I was
looking for some way to use axis() to draw an unlabeled axis and then
put in text labels after that. However, I can't understand how to get
the values of the tick marks placed by axis from the figure in order to
place text by some tick marks.
2. Some students make postscript figures that fit "just right" into
LaTeX documents, while some make figures that have huge,
inches-and-inches of margins around the outside. I'm not watching how
they make these figures all the time, but I think I'm figuring out the
cause of big margins.
Is this right: the margins issue relates to the use of the onefile=F
option in a dev.copy command? I think the figures turn out properly
with this:
dev.copy(postscript,file="myfig.eps",height=5,width=5,horizontal=F,onefile=F)
dev.off()
If you mistakenly omit the onefile=F option, the margins stretch out
from the center where the figure is placed to the edges of an entire
sheet of paper. In other words, the default settings for margins in
inches that I see in output from
> par()
...
$mai
[1] 1.0066821 0.8092935 0.8092935 0.4145162
...
have no effect if a person forgets the onefile=F option. We fiddled a
while, trying to force margins down, par(mai=c(0,0,0,0)), but no matter
what we did, the figures still had the massive margins.
We don't have these margin problems with other devices. Margins in jpeg
or png pictures are sized appropriately.
--
Paul E. Johnson email: pauljohn at ku.edu
Dept. of Political Science http://lark.cc.ku.edu/~pauljohn
1541 Lilac Lane, Rm 504
University of Kansas Office: (785) 864-9086
Lawrence, Kansas 66044-3177 FAX: (785) 864-5700
Greg Snow
2005-Sep-21 22:08 UTC
[R] controlling usage of digits & scientific notation in R plots; postscript margins
One option is to use the 'scipen' option. see ?options and look for scipen. An example: options(scipen=3) # now do your plot and see what happens (try bigger numbers if 3 doesn't change anything). you can also specify different axis locations and labels to do it by hand: plot( (1:10)*1e+5, 1:10, xaxt='n') axis(1, at=(1:10)*1e+5, labels=paste( 1:10, "*10^5", sep='')) also look at ?plotmath if you want actual superscripts. Hope this helps, Greg Snow, Ph.D. Statistical Data Center, LDS Hospital Intermountain Health Care greg.snow at ihc.com (801) 408-8111>>> Paul Johnson <pauljohn at ku.edu> 09/21/05 03:42PM >>>Dear R users: I assigned students to make some graphs and I'm having trouble answering some questions that they have. We are all working on R 2.1 on Fedora Core Linux 4 systems. 1. In the plot, the axis is not labeled by "numbers", but rather scientific notation like "-2e+08" or such. We realize that means -200,000,000. We want to beautify the plot. We would rather just print out the whole, big number. But if we can't have that, we would like something more beautiful and mathematical, such as 8 -2.0x10 Once the numbers are big, R automagically switches to scientific notation, and turning the values horizontal does not help on the y axis. Example: x <- rnorm(100,mean=100000,sd=100000000) y <- rnorm(100,mean=10000000000,sd=100000000000) plot(x,y,axes=F) axis(2,las=2) #turns y axis numbers perpendicular to y axis axis(1) # x axis I realize a person could just divide all numbers by 1 million and then have a graph with -200 and an annotation (in millions). On the x axis, we'd even settle to just label the two outermost points with full numerical values, and then have only ticks between. I was looking for some way to use axis() to draw an unlabeled axis and then put in text labels after that. However, I can't understand how to get the values of the tick marks placed by axis from the figure in order to place text by some tick marks. 2. Some students make postscript figures that fit "just right" into LaTeX documents, while some make figures that have huge, inches-and-inches of margins around the outside. I'm not watching how they make these figures all the time, but I think I'm figuring out the cause of big margins. Is this right: the margins issue relates to the use of the onefile=F option in a dev.copy command? I think the figures turn out properly with this: dev.copy(postscript,file="myfig.eps",height=5,width=5,horizontal=F,onefile=F) dev.off() If you mistakenly omit the onefile=F option, the margins stretch out from the center where the figure is placed to the edges of an entire sheet of paper. In other words, the default settings for margins in inches that I see in output from > par() ... $mai [1] 1.0066821 0.8092935 0.8092935 0.4145162 ... have no effect if a person forgets the onefile=F option. We fiddled a while, trying to force margins down, par(mai=c(0,0,0,0)), but no matter what we did, the figures still had the massive margins. We don't have these margin problems with other devices. Margins in jpeg or png pictures are sized appropriately. -- Paul E. Johnson email: pauljohn at ku.edu Dept. of Political Science http://lark.cc.ku.edu/~pauljohn 1541 Lilac Lane, Rm 504 University of Kansas Office: (785) 864-9086 Lawrence, Kansas 66044-3177 FAX: (785) 864-5700 ______________________________________________ R-help at stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html
Prof Brian Ripley
2005-Sep-22 07:33 UTC
[R] controlling usage of digits & scientific notation in R plots; postscript margins
Concise answers to verbose questions: 1) Use options(scipen) (and probably change the margin sizes). Or something like options(scipen=10) par(mar=c(5,8,4,2)+0.1) plot(x, y, axes=FALSE) axis(2, las=2) axis(1, labels=FALSE) axis(1, at = c(-2e8, 2e8), labels = expression(-2 %*% 10^8, 2 %*% 10^8)) 2) This is a function of the including application, in particular the macro package used in latex. Some use the bounding box (rather than the paper size) for files marked with EPSF in the header, which is done if onefile=FALSE (sic). For others you need paper="special" as well. Hence the advice in ?postscript. Just follow it! On Wed, 21 Sep 2005, Paul Johnson wrote:> Dear R users: > > I assigned students to make some graphs and I'm having trouble answering > some questions that they have. We are all working on R 2.1 on Fedora > Core Linux 4 systems.There is no `R 2.1': please do study the posting guide.> 1. In the plot, the axis is not labeled by "numbers", but rather > scientific notation like "-2e+08" or such. We realize that means > -200,000,000. We want to beautify the plot. We would rather just print > out the whole, big number. But if we can't have that, we would like > something more beautiful and mathematical, such as > 8 > -2.0x10 > > Once the numbers are big, R automagically switches to scientific > notation, and turning the values horizontal does not help on the y axis. > > Example: > > x <- rnorm(100,mean=100000,sd=100000000) > y <- rnorm(100,mean=10000000000,sd=100000000000) > plot(x,y,axes=F) > axis(2,las=2) #turns y axis numbers perpendicular to y axis > axis(1) # x axis > > I realize a person could just divide all numbers by 1 million and then > have a graph with -200 and an annotation (in millions). > > On the x axis, we'd even settle to just label the two outermost points > with full numerical values, and then have only ticks between. I was > looking for some way to use axis() to draw an unlabeled axis and then > put in text labels after that. However, I can't understand how to get > the values of the tick marks placed by axis from the figure in order to > place text by some tick marks. > > 2. Some students make postscript figures that fit "just right" into > LaTeX documents, while some make figures that have huge, > inches-and-inches of margins around the outside. I'm not watching how > they make these figures all the time, but I think I'm figuring out the > cause of big margins. > > Is this right: the margins issue relates to the use of the onefile=F > option in a dev.copy command? I think the figures turn out properly > with this: > > dev.copy(postscript,file="myfig.eps",height=5,width=5,horizontal=F,onefile=F) > dev.off() > > If you mistakenly omit the onefile=F option, the margins stretch out > from the center where the figure is placed to the edges of an entire > sheet of paper. In other words, the default settings for margins in > inches that I see in output from > > > par() > ... > $mai > [1] 1.0066821 0.8092935 0.8092935 0.4145162 > ... > > have no effect if a person forgets the onefile=F option. We fiddled a > while, trying to force margins down, par(mai=c(0,0,0,0)), but no matter > what we did, the figures still had the massive margins. > > We don't have these margin problems with other devices. Margins in jpeg > or png pictures are sized appropriately. > > > -- > Paul E. Johnson email: pauljohn at ku.edu > Dept. of Political Science http://lark.cc.ku.edu/~pauljohn > 1541 Lilac Lane, Rm 504 > University of Kansas Office: (785) 864-9086 > Lawrence, Kansas 66044-3177 FAX: (785) 864-5700 > > ______________________________________________ > R-help at stat.math.ethz.ch mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide! http://www.R-project.org/posting-guide.html >-- Brian D. Ripley, ripley at stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UK Fax: +44 1865 272595