Unsupervised Learning — Clustering, Anomalies, and Dimensionality Reduction

When the data has no target column, you're in unsupervised territory. Three things you can do with it:

1. Cluster to find segments / groups
2. Detect anomalies to find outliers / fraud / failures
3. Reduce dimensions to visualize, denoise, or compress

Each comes with its own gotchas. The biggest one: there's no single "accuracy" metric, so you have to be deliberate about how you evaluate.

When to use this skill