The best explanation of PCA and PLS-DA comes from the group where it originated - Svante Wold's group at Umea university (see wikipedia definition of chemometrics for potted version of the history)
Cross-Validatory Estimation of the Number of Components in Factor and Principal Components Models
Svante WoldTechnometricsVol. 20, No. 4, Part 1 (Nov., 1978), pp. 397-405Published by: Taylor & Francis, Ltd. on behalf of American Statistical Association and American Society for QualityDOI: 10.2307/1267639
or
A chemometrics toolbox based on projections and latent variables
Lennart Eriksson Johan Trygg Svante WoldFirst published: 13 January 2014 https://doi.org/10.1002/cem.2581
There are some good articles such as
https://www.cs.princeton.edu/picasso/mats/PCA-Tutorial-Intuition_jp.pdf that present PCA from both geometric and matrix algebra aspects.
There are plenty of other visual descriptions and youtube tutorials (see below)
Pearson published a principal-axis approach in 1901 in UK, and Hotelling further developed it in 1933 in USA. These are the historical references (when PCA is considered created). I can provide you details about the references if you want (just send me an email). However, there were later articles that explained PCA aiming to concrete purposes or in a more pedagogical way. If you want a reference to understand and see how to use PCA, I would recommend Carey 1975 or Jolliffe 2002 (the latter is a book entitled Principal Component Analysis, the author is I.T. Jolliffe, and the book is part of the Springer Series in Statistics), as well as some application articles to see some examples. On the other hand, most of chemometrics books have an introduction to PCA including examples.