Unsupervised Machine Learning Tactics
For Ideal Customer Profile Segmentation

Twin Chutes Analytics Data & AI Engineer Pete Osorio outlines some quick strategies for visualizing unsupervised clustering algorithms operating in higher dimensional space in 2-dimensions to aid commonly used analytical indicators, such as Davies-Bouldin and Silhouette scores.

Marketing departments frequently identify ideal customer profiles for targeted campaigns, creatives, and digital ads. These are typically based on learned heuristics from organizational leadership with domain expertise, however there are opportunities to put additional analytical rigor into these decisions.