This article is relatively popular here (considering the topic). Two previous discussions about it: From 2015: https://news.ycombinator.com/item?id=9040266 htt

This article is relatively popular here (considering the topic). Two previous discussions about it: From 2015: https://news.ycombinator.com/item?id=9040266 https://news.ycombinator.com/item?id=9040266. From 2017: https://news.ycombinator.com/item?id=14405665 https://news.ycombinator.com/item?id=14405665. A more recent approach to visualizing high-dimensional data is the t-SNE algorithm, which I normally use together with PCA when exploring big data sets. If you're interested in the differences between both methods, here's a really good answer: https://stats.stackexchange.com/a/249520 https://stats.stackexchange.com/a/249520.