coordinate vs contributions for the variables in PCA using FactoMineR

Question

[QX]

Hi all,

I am using PCA function from FactoMineR package, which they return these 2 dataframes:

"$var$coord" "coord. for the variables"

"$var$contrib" "contributions of the variables"

can anyone help me explain the difference between these two factors? If I want to check what variables affect the separations on PC1 of PCA plot by ranking, which $var would be the right choice?

Answer 1

[Bastien Hervé]

The answers to your questions are already in your post.

$var$coord gives you the coordinates of your variables on each principal component

$var$contrib gives you the contributions of each variable to each principal component (how much they contribute)

If I want to check what variables affect the separations on PC1 of PCA plot by ranking, which $var would be the right choice?

You use $var$contrib

Get familiar with Individuals/Variables, eigenvalues, eigenvectors, loadings, correlation

http://www.sthda.com/english/wiki/wiki.php?id_contents=7851

https://stats.stackexchange.com/questions/143905/loadings-vs-eigenvectors-in-pca-when-to-use-one-or-another

[QX]

thank you!

[QX]

Hi, one additional question, for this image:

if I want to know which variables that separate the yellow and blue dots, which $var should I use? I guess the $coord, right? as it represents the coordinate on the PC1

[Bastien Hervé]

It is the combinaison of your variable that separate yellow and blue dots on PC1/PC2. If you want to know which one drive this behaviour the most you still have to look at $var$contrib

$coord only describe the localization of your dots on your screen