K-Means Clustering - k = 2–6, three initialisations and a WCSS chart

Data Science & Big Data Analytics • ITec4121 • Wolkite University • Online-First

Run k-means by hand: Step does one assignment + centroid update, Auto runs to convergence. Choose any k from 2 to 6, three initialisation strategies (bbox spread, k-means++ or random points), the point count for the random set (16/32/40), three extra datasets (oranges, expanded iris, two moons), and an auto-run speed. A live WCSS-by-iteration chart shows what the algorithm is really minimising.