Unsupervised learning works with data that has no labels. There is no "right answer" to learn from, so the goal is to find structure on your own: grouping similar items together (clustering) or squeezing many features down to the few that matter most (dimensionality reduction). It's how you segment customers, spot unusual patterns, or make high-dimensional data possible to visualize.
The first three lessons are short, visual explanations of k-means clustering, hierarchical clustering, and PCA, with no code, just intuition. The last video puts k-means into practice with scikit-learn so you can see it run on real data.
Watch the lessons in order. If a video runs long, feel free to treat it as a reference you dip back into later rather than something to finish in one sitting.
How k-means groups unlabeled data: pick K, assign points to the nearest center, update the centers, and repeat, plus the elbow method for choosing K.
Video 8minBuilding a tree of nested clusters from the bottom up, how distance and linkage choices change the result, and how to read a dendrogram and heatmap.
Video 11minThe core idea of principal component analysis: reduce many correlated features to a few components that keep most of the information in your data.
Video 6minTech With Tim shows how to run k-means in Python with scikit-learn: loading a dataset, fitting the model, and evaluating the clusters it finds.
Video 12min