Laydi is an acronym for look at your data interactively, which is what the program is aimed at. It is a lightweight data analysis program for bilinear modeling (PCA and PLS) with a strong focus on interactive use. Laydi is released under the GNU GPL and the latest development snapshot can be downloaded from https://git.pvv.ntnu.no/Projects/laydi.git

Features

Principal Component Analysis (PCA)
Partial Least Squares Regression (PLS)
L-shaped PLS regression (L-PLS)
Easy mapping of variables between plots, selections in one plot propagates to other plots.

Nonfeatures

Does not import arbitrary files. Files must be prepared in a (simple) file format prior to import.
Saving and loading of projects is not implemented. (Datasets can be saved and loaded, though, and plots can be exported)
Not very stable

Installation requirements

Laydi currently requires the following extra packages, available from apt on Debian and Ubuntu.

python2.4 or python2.5
python-glade2
python-gnome2
python-gtk2
python-matplotlib
python-scipy
python-numpy

Partially needed

python-networkx
python-pygraphviz

Download laydi

Laydi is not debianized. To download it, use the clone the git repo.

git clone https://git.pvv.ntnu.no/Projects/laydi.git

User documentation

Frequently Asked Questions
Laydi help (the same as available through the help menu in the application.)
Terminology

Developer documentation

Developer tips and tricks