The following are links to sites with (mostly) free general purpose statistical
R. A free version of the
S/S-plus family of statistical programming languages. The contributed
code at the Comprehensive R Archive Network (CRAN) provides packages for an
enormous number of statistical procedures.
(emacs speaks statistics) An (X)emacs interface for interacting with
SAS, S-plus/R, xlispstat, etc.
xplore. Started with a very broad
set of smoothing procedures and includes now client/server
architecture for web computing. Not exactly free (no source code
etc.), but a limited version (allowing for up to 1000 obsevations only
and with fixed expiration date. The Linux version has no expiration
date.) is available.
Ox. A powerful
matrix language, batch oriented (no source code provided).
BUGS. Bayesian inference
Using Gibbs Sampling (no source code provided).
Data and Code Archives
StatLib. Code for several software
packages, published data sets from many journals, and published or else
Statistical Reference Datasets from NIST.
Its aim is to increase the accuracy of statistical software by providing
standard test datasets.
(a collection from newspapers and other sources, for teaching statistics).
Data and Story Library, DASL.
A project similar in scope to CHANCE.
The German Socio-Economic Panel (GSOEP). SOEPINFO-WWW is an
interactive system with extensive search capabilities that provides
you with detailed information on the variables in the SOEP dataset. By
interactively creating lists of variables in SOEPINFO, the user can
then output frequencies information, item correspondence, and even
generated SPSS, SAS, TDA (6.1) and Stata (7.0) command files. SOEPINFO
should be your first step in starting any new SOEP project !
A collection of links to further data archives is here.
The textbooks currently avaiblable on the net are very heterogeneous in
scope, completeness, and use of web techniques. They may incorporate just
plain texts translated to HTML or experiment with interactive graphics
and web computing. Some do require a local installation of plug-ins or
of statistical software. Nevertheless, they provide examples of what may
be possible through the internet.