Correspondence Analysis (CA) with python

Implementation through an example

File to download to obtain the code and example data files

Loading the compressed code and example data files [folder: "CA_py.zip"]

The comprehensive Correspondence Analysis (CA) program presented here is designed in the spirit of exploratory data analysis. We thus insist on the fundamental role, for the interpretations, of additional (or supplementary) variables (numerical or nominal) and additional (or supplementary) individuals.

This stand-alone program does not refer to existing CA programs. It only uses the three classic python libraries NUMPY (for vector and matrix computation), PANDAS (for managing data tables) and MATPLOTLIB (for graphics)..

  CA: Example and listing of the commented code

This small introductory example is commented on in detail in Chapter 3 of the book "Exploring Textual Data" (Lebart, Salem and Berry, Kluwer Academic Publishers, 1998). An adapted excerpt of chapter 3 containing the example can be downloaded from: [ Exploring Textual Data, excerpt from Chapter3 ]. The whole book itself is not yet in the public domain. The same example data set was already used in chapter 3 of the book "Statistique Textuelle" (in French) (Ludovic Lebart and André Salem, Dunod, 1994,Paris, for the first printing). This book is freely downloadable on this website: [Statistique Textuelle] .

Click below to get the presentation of CA from an example

Commented examples and python program