S-PLUS Applications and Links

S-PLUS is a powerful interactive package for statistical and graphical analysis. It is a combination of an object-oriented programming language, powerful graphics capabilities, and a large number of pre-programmed statistical routines. The programmability of S-PLUS gives it extreme flexibility, which is why it is very popular with statistical researchers. However, this flexibility can be intimidating to the first time user.

While this page cannot provide a complete tutorial on using S-PLUS, it does cover the basics of starting S-PLUS in a Unix environment, such as VET, and a quick-start guide to on-line help in S-PLUS. Also included are pointers to more information, on the web and elsewhere, about S-PLUS.

One particular web resource will be pointed out here: Notes on S-PLUS by Bill Venables and David Smith is a Postscript file containing an 87 page tutorial on S-PLUS. This is probably a good place for the novice to start.

Using S-PLUS on VET

The normal mode for S-PLUS is an interactive, or command-line format, where each command is typed and the user waits for the results. There is a ``batch'' mode, where a file containing a list of commands is executed ``in the background'', but this is not covered here. More information about this can be obtained by examining the help inside S-PLUS on ``BATCH'' (see below for more information about the S-PLUS online help).

To start the S-PLUS interactively on VET, the simply type

% Splus
(the ``%'' is the shell prompt, thus it is not typed). The first time S-PLUS is used, it creates a directory named ``.Data'' in the your home directory. S-PLUS stores all of the data objects created by you in this directory, as well as some information used by S-PLUS itself. After this creation and some other initialization, the S-PLUS prompt, a greater-than (>) sign, appears. This means that S-PLUS is ready for input.

To get graphics in S-PLUS, a ``device'' must be initialized. In X (X-windows), the easiest device to use is the Motif device. To start this device, simply type the command

motif()
after S-PLUS has started. This command runs the function motif, which opens a graphics window and instructs S-PLUS to do all its graphing in this window.

S-PLUS has extensive help on all of the built-in functions and datasets. In X-windows, a window-based help utility is started by issuing the command

help.start()
The window that pops up has three sub-windows: a list of functions, a list of categories, and a topic entry window. These operate as follows: These three capabilities make it easy to find help in S-PLUS. Furthermore, if there is a specific function that you need help with (and you already know the name of the function exactly) you may use the help function at the S-PLUS command line (not in the help window). For example, if you know you need help with lsfit, you can simply type the command
help("lsfit")
This will immediately display the help for that function.

Electronic Resources on S-PLUS

Because of its heavy use in academic statistics community, there are many web-based resources on S-PLUS. A small segment is given here, highlighting the sites that are deemed most useful.


The MathSoft S-PLUS Site

MathSoft is the developer and distributor of S-PLUS, thus this is the ``home base'' for S-PLUS.

S Frequently Asked Questions

A list of Frequently Asked Questions about S and S-PLUS, gleaned from the S-news mailing list (see below).

S-news mailing list

Sending a one-line message to s-news-request@utstat.toronto.edu saying ``subscribe'' will get you a subscription to this rather active (about 20 to 30 messages a day) mailing list all devoted to S-PLUS.

StatLib S Archive

This is a repository of extensions to S-PLUS made by the general user community. This can sometimes be helpful, when a particular analysis is needed but is not already in S-PLUS.

Notes on S-PLUS, by B. Venables and D. Smith

This is a 744K Postscript file containing a short-course on S-PLUS, including exercises. A very good place for the novice to start, since it is freely available over the web. The Postscript file can be viewed in Ghostview or printed on a Postscript printer (there are 87 pages in all).

Statistical Computing Tools and Documents

This page, maintained by F. E. Harrell, a biostatistician at the University of Virginia, contains much information on two extensions to S-PLUS, the Hmisc library and the Design library. These require installation, which can be non-trivial. However, of particular interest is the document An Introduction to S-Plus and the Hmisc and Design Libraries which includes introductory information on S-PLUS, including a comparison between S-PLUS and SAS.

Books on S-PLUS

While not an exhaustive list of books on S-PLUS, this list covers the main introductory texts on S-PLUS. Each book has a slightly different approach, often relating to the approach to statistics that the authors take.


The New S Language, by R. A. Becker, J. M. Chambers and A. R. Wilks, 1988.

This is known as ``The Blue Book'' in the S-PLUS community. It is now a little dated, but most of the information is still relevant. Note that an earlier book (1984) by Becker and Chambers is out-of-date and should not be used.

Statistical Models in S, by J. M. Chambers and T. J. Hastie, 1992.

This book takes up where the The New S Language leaves off, and deals with S-PLUS from a modeling standpoint. It is a good reference for the more advanced features of S-PLUS.

An Introduction to S and S-plus, by P. Spector, 1994.

This is an introductory book on the basics of S-Plus.

Modern Applied Statistics with S-PLUS, 2 ed., by W. N. Venables and B. D. Ripley, 1997.

This recently (July 1997) published second edition is very popular. It covers all the basics of S-PLUS from an applied standpoint, but also is designed as a text-book to teach modern statistical methods. This book has its own web page, which contains more detailed information.

Statistical Consulting Lab Links

 

Mission of the Statistical Consulting Lab

Recent Projects

Vet and Unix Information

Statistical Software

Data Resources

SCL Tips

Requests for Help and Discussion

  • Software Related
  • Research Related

Fall Schedule for Lunch Seminars

Statistical Links

Home to the Statistical Consulting Lab Page