About this module
Next we connect variables without claiming every relationship is causal.
Paired observations become scatter plots, centered products, ranks, and a fitted line.
We will inspect residuals and group composition, then test the boundary between a useful predictive summary and a justified explanation of an intervention.
Lessons
- Lesson 81
Scatter Plots, Association, and Nonlinear Patterns
Looking at paired observations before compressing them
2:34 - Lesson 82
Sample Covariance Calculation and Interpretation
Constructing a covariance rather than trusting a matrix cell
2:27 - Lesson 83
Pearson Correlation Calculation and Interpretation
Normalizing linear association without claiming independence
2:33 - Lesson 84
Spearman Rank Correlation and Kendall Tau
Comparing monotone rankings and pair ordering
2:40 - Lesson 85
Correlation, Causation, Confounding, and Spurious Relationships
A pooled recovery-rate comparison that reverses within difficulty groups
2:35 - Lesson 86
Simple Ordinary Least Squares Regression
Fitting a line to paired operational measurements
2:36 - Lesson 87
Intercepts, Slopes, Coefficients, and Predictions
Interpreting a coefficient before using it for extrapolation
2:37 - Lesson 88
Residuals, MAE, MSE, and RMSE
Choosing an error metric that reflects the cost of mistakes
2:34 - Lesson 89
R-Squared and Adjusted R-Squared
Comparing a regression with a mean-only baseline
2:33 - Lesson 90
Regression Assumptions, Heteroskedasticity, and Multicollinearity
Diagnosing unstable uncertainty and redundant predictors
2:56