Multivariate and Clustering

Start here when your question is exploratory and involves many variables, grouped profiles, reduced-space views, or candidate cluster structure rather than single distributions or simple two-variable relationships.

Overview

  • Use this branch when your question is exploratory and involves many variables, grouped profiles, reduced-space views, or candidate cluster structure.
  • If you mainly need a single-variable distribution or grouped comparison view, use Distribution and Group Comparison; if you mainly need a two-variable relationship or agreement view, use Association and Regression.
  • This branch covers exploratory multivariate and clustering displays, not proof of separation, confirmation of latent structure, or automatic discovery results.
  • Some pages are also narrower than their titles suggest, so page-level support boundaries matter.

Child pages

Start here

  1. Start with Heatmap when you want to scan many values for matrix-like pattern candidates.
  2. Use Parallel Coordinates when each observation has several variables and the profile across axes matters.
  3. Use PCA Biplot when you need a cautious reduced-space view of multivariable structure, not proof that groups are truly separated.
  4. Move to K-means Clustering Plot or Hierarchical Clustering Heatmap only when clustering itself is part of the exploratory question.
  5. Use Confidence Ellipse or Convex Hull when you want grouped scatter overlays, not standalone proof-like summaries.
  • Related links are registered in frontmatter and rendered below this stub.