R help - Dec 2001 - Graphics with moderately large amounts of data

Hi,

A major attraction to R and to S-plus are the graphics. (Up to now my
experience is with 
STATA and SAS.) Most of the graphical examples that I have seen in the
documentation 
are for relatively small size data sets. I am working with a moderately
large data set -- 
the order of magnitude is 180,000 observations by 50 variables. There seem
to be standard 
problems that I keep bumping into in the graphics: eg. the graphics work
hard to accomodate 
outliers leaving the main action area a thick cloud, very slow operations by
R, etc. I have 
been doing some obvious things to deal with these issues, eg. trimming,
restricting attention 
to data subsamples, etc. But these must be pretty standard issues. I would
like to take 
advantage of what is already known. What should I be reading that explains
how best to
do graphics with a somewhat larger data set? (Pointers to an appropriate FAQ
would be 
great since I have looked but not managed to find it.)

Thanks in advance for any advice.

Thanks.
Murray.

Murray Z. Frank
Faculty of Commerce
University of British Columbia
Vancouver, B.C.
Canada V6T 1Z2

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

>>>>> "Frank," == Frank, Murray <murray.frank at
commerce.ubc.ca> writes:
    > I am working with a moderately large data set -- the order
    > of magnitude is 180,000 observations by 50 variables. There
    > seem to be standard problems that I keep bumping into in
    > the graphics: eg. the graphics work hard to accomodate
    > outliers leaving the main action area a thick cloud, very
    > slow operations by R, etc. I have been doing some obvious
    > things to deal with these issues, eg. trimming, restricting
    > attention to data subsamples, etc.

I don't know of an appropriate FAQ to refer you to.  The issue of
outliers can be dealt with by setting the plotting limits with
xlim and ylim.  This might be faster than trimming the data set.

As far as graphics speed goes, a non-R possibility for faster
plotting of large data sets is root (http://root.cern.ch).
(Caveat: I've used root for large data sets (~10^6 measurements
of many parameters) and I have not used R for data sets larger
than several thousand measurements of a dozen or so parameters.
I have not made a direct comparison with large data sets - YMMV.)

Mike
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._