[R] How to get correct proportions/bounding box for latex figure?

[Duncan Murdoch]

context grey wrote:
> That's approximately right, but the individual
> scatterplots
> are slightly stretched horizontally.   
> 
> Is there not any way to have the plots have true 1:1
> aspect ratio
> (given that the range of the data is the same on both
> axes)
> and still get a bounding box?. 

What do you mean by 1:1 aspect ratio?  Do you mean the physical 
measurement of one vertical unit should match the physical measurement 
of one horizontal unit?  If so, use

aspect="iso"

in your calls to xyplot.  If you mean that the boxes should be square, use

aspect=1.

(I hope I've got these right; I don't use lattice much.)

In both cases, R is not that good at calculating the bounding box:  it 
will often include a lot of white space.  (In this case it looks as 
though it is believing the values passed in the trellis.device call, but 
with other args to the postscript device it will do other things.) I 
generally use GSview's PS to EPS function to recalculate it when this 
sort of thing matters.

Duncan Murdoch

   And, without getting
> out
> a ruler and manually calculating width/height values?
> 
> (I'm guessing you arrived at width=12,height=4 just
> by "eyeballing" it.?)
> 
> --- Martin Sandiford <ms at mcdev.com.au> wrote:
> 
> 
>>Does this do what you want?
>>
>>library(lattice)
>>
>>rand1 <- rnorm(50)
>>rand2 <- rnorm(50)
>>theplot <- xyplot(rand1 ~ rand2, xlab="x axis",
>>ylab="y axis")
>>
>>thefile <- "plotproblem.eps"
>>trellis.device(postscript, file=thefile, color=F,
>>horizontal=FALSE, width=12, height=4,
>>paper="special")
>>print(theplot, split=c(1,1,3,1), more=T)
>>print(theplot, split=c(2,1,3,1), more=T)
>>print(theplot, split=c(3,1,3,1), more=F)
>>dev.off()
>>
>>(Paper size and type specified in trellis.device
>>call).
>>
>>
>>Martin
>>
>>On 15/03/2006, at 2:44 PM, context grey wrote:
>>
>>
>>>
>>>--- Duncan Murdoch <murdoch at stats.uwo.ca> wrote:
>>>
>>>>The R graphics model is that the drawing surface
>>
>>is
>>
>>>>established first,
>>>>then the things you draw are adjusted to fit in
>>
>>it.
>>
>>>>R won't change the
>>>>shape of the display because you are drawing more
>>>>things on it.
>>>
>>>Thanks, this comment clarifies things somewhat.
>>>Though I find it
>>>an odd design choice for R, since it seems to
>>
>>entail
>>
>>>then that
>>>the user has figure out the aspect ratio of the
>>>resulting plot,
>>>something that R could easily keep track of as it
>>
>>is
>>
>>>drawing.
>>>
>>>Here's example code, producing 3 scatterplots
>>>side-by-side
>>>(here reusing the same plot for simplicity).
>>>
>>>What's desired is that the individual scatterplots
>>>have the
>>>natural aspect, e.g. square, with the axis units
>>
>>being
>>
>>>the
>>>same for X, Y.   And to do this while producing a
>>>correct
>>>bounding box in the .eps file.
>>>
>>>As it stands the example code produces a correct
>>>bounding box,
>>>but the scatterplots are too stretched to be
>>
>>usable.
>>
>>>Inserting
>>>aspect=1/1 in the xyplot() seems to cause the
>>
>>bounding
>>
>>>box
>>>to be  incorrect.
>>>
>>>------------
>>>
>>>library(lattice)
>>>
>>>rand1 <- rnorm(50)
>>>rand2 <- rnorm(50)
>>>theplot <- xyplot(rand1 ~ rand2, xlab="x axis",
>>>ylab="y axis")
>>>
>>>thefile <- "plotproblem.eps"
>>>trellis.device(postscript, file=thefile, color=F,
>>>horizontal=FALSE)
>>>print(theplot, split=c(1,1,3,1), more=T)
>>>print(theplot, split=c(2,1,3,1), more=T)
>>>print(theplot, split=c(3,1,3,1), more=F)
>>>dev.off()
>>>
>>>-----------------------
>>>
>>>
>>>
>>>>I don't think I understand exactly what you want
>>
>>to
>>
>>>>achieve; sample code
>>>>that produces something close would be helpful
>>
>>(even
>>
>>>>if it comes out the
>>>>wrong shape).
>>>
>>>______________________________________________
>>>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
>>>
>>
>>
> 
> ______________________________________________
> 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