[R] New package announcement: R2STATS, a GUI for fitting GLM and GLMM

[Yvonnick Noel]

Dear R-users,

I wanted to inform you that a new package called R2STATS is available, 
as a graphical front-end for the glm() and glmer() functions.

The GUI is based on the RGTk2 and gWidgets packages by Michael Lawrence 
and John Verzani, and so requires that the GTK+ library be installed 
first on your system. This is done "automagically" when installing the 
RGtk2 package (or the script mentioned below). It also use the 
RGtk2Extras by Tom Taverner, to provide editable grids for data frames.

This GUI is intended to provide an easy way to fit and compare GLM and 
GLMM models. The GLMM part is based on Douglas Bates' lme4 package and 
the glmer() function. Automatic plots are also drawn for every model, 
and you can switch from one plot to the other by just clicking on the 
model name. I found this feature quite useful when teaching: It helps 
students to get an immediate understanding of differences between models.

Note that this GUI is left (deliberately) simple and is not intended to 
provide a full-featured GUI (please consider using Rcmdr instead for a 
far more advanced GUI). But it tries to do well the one and only thing 
it was designed to do: Fitting and comparing models. Note that most 
standard statistical tests may well be presented as a simple comparison 
between GLMs and this is the way I go with my students here. This allows 
an integrated presentation for almost all common (and simple) situations 
in social sciences.

More information is available on my webpage : 
http://yvonnick.noel.free.fr/r2stats [in French for the moment, although 
the package is in English].

Installing the package is done from a temporary repository:

install.packages("R2STATS",repos="http://yvonnick.noel.free.fr/cran",dep=TRUE)

if you already have a recent version of GTK+ and RGtk2 installed, or by:

source("http://yvonnick.noel.free.fr/r2stats/installwin.R")

for an automatic script that download and install everything. I will 
submit it to CRAN as soon as I have fixed some minor issues with R-devel 
(but the package works flawlessly with the current R-2.13.2).

Any comment welcome. Also, if you are willing to contribute a 
translation into your language, please let me know.

Best,

Yvonnick Noel, PhD.
University of Brittany
Rennes, France