The PowerPoint in this repo is a presentation I gave at the Richmond Data Science Community Meetup (see https://www.meetup.com/Richmond-Data-Science-Community-Meetup/). The slides cover a few dimensionality reduction techniques, starting with an overiew of the purpose of dimensionality reduction and PCA. The rest of the material goes into Independent Component Analysis (ICA), Random Projections, image compression, and NMF clustering.
In my blog, I have an article going into more detail about doing Indepedent Component Analysis in Python (check out this link: http://theautomatic.net/2018/06/23/ica-on-images-with-python/). This extends upon an example shown in the slides.