19  Correlation Isn’t Dependence

Builds covariance and correlation from scratch, then constructs a deliberately adversarial dataset where correlation is near zero but the variables are strongly dependent in a nonlinear way — proving by counterexample that correlation is a summary of linear association only.