A shop has a Dataset of six customers. Each row has two numbers: visits per month and average spend. Nobody has written a label such as "bargain hunter" or "loyal regular" next to any row.
Earlier lessons trained on rows that came with an answer. This table has none. The plot shows the six customers as plain grey dots, and the question is what, if anything, a program can find in them without being told what to look for.
💡Good to know
Unsupervised learning means the data has no target column. The model looks for structure in the inputs alone.