Stat 850 Intro to Computing

Here are some brief hints about using computers for this course. It is strongly recommended that you learn how to use the World Wide Web, as that is an easy way to pass information. There is introductory documentation on both the sas and S statistical languages at our web pages, under Computing Facilities. See also the UNIX Guide. Look under Statistical Software Packages & Languages.

World Wide Web

The Web is a world-wide information architecture, which includes gopher and ftp as "special cases". Computers all over are hooked in. In order to access it, you need to use a "browser", which is just a piece of computer software. Our system has four browsers you can choose from:

Mosaic http://www.stat.wisc.edu/: the first web browser, developed by NCSA at U IL
/usr/unsup/netscape/netscape http://www.stat.wisc.edu/: only availabe on fisher and pearson
/usr/psup/lynx/bin/lynx http://www.stat.wisc.edu/: text browser, much like gopher
w3 (only from within emacs): text browser (see http://www.stat.wisc.edu/emacs.html)

Mosaic and netscape are graphics-based browsers, while lynx and w3 are text-based browsers. The "address" http://www.stat.wisc.edu/ is a universal resource locator (URL) for the U WI Statistics Department's "Home Page" on the web. You can find out more about these browsers and our facilities by "visiting" this home page, and clicking on highlighted text. It's fun!

Introductions to SAS Language

Here is a quick start for sas. Create a directory and move to it:

% mkdir st850
% cd st850

Copy the Milliken & Johnson pulse data and SAS files from the public data area /p/stat/Data/MJ:

% cp /p/stat/Data/MJ/pulse.dat .
% cp /p/stat/Data/MJ/pulse.sas .
% cp /p/stat/Data/MJ/pulse.log yandell.log
% cp /p/stat/Data/MJ/pulse.lst yandell.lst

Notice that I changed the name so that you won't overwrite them when you run SAS on your own:

% sas pulse.sas

Be sure to always check the log file!

Introductions to S (or Splus) Language

Here is a quick start for S or Splus. Suppose you already have the pulse data used above. Copy the Milliken & Johnson S source and listings files from the public data area /p/stat/Data/MJ:

% cp /p/stat/Data/MJ/pulse.s .
% cp /p/stat/Data/MJ/pulse.sink .

(If you want to print pulse.sink, first edit out the printer() plots to save paper!) Try to reproduce my listings in your own session. Type S (on bayes) or Splus (on atlas) to begin:

% S
Working data will be in /u/stat/faculty/yandell/.Data
>

If this is your first time using S, you will be asked to create a directory with the name ".Data". Do it. Now you can source the S commands in the file pulse.s,

> source("pulse.s")

Then proceed through the commands in the file pulse.sink -- that is, the lines which begin with the less-than prompt (> ). Turn on a graphics device such as x11() or X11() before the plot:

> pulse.fit
> summary(pulse.fit)
> names(pulse.fit)
> anova(pulse.fit)
> x11()
> plot(pulse.fit)

The screen should give you a GO? prompt after the first plot, indicating that further plots are possible by hitting the return key. As indicated in the web Introduction to S, you can use the postscript() command to save plots as PostScript for later printing:

> postscript("pulse.ps")
> plot(pulse.fit)
> q()

It is a good idea to preview plots with ghostview before sending it to the printer:

% ghostview pulse.ps
% lpr pulse.ps

Last modified: Sun Jan 28 14:29:11 1996 by Brian Yandell yandell@stat.wisc.edu