Correlation Matrices

Correlation matrices show the pairwise relationships between multiple variables at once. They're essential for exploratory data analysis and feature selection.

Basic Correlation Matrix

Lower Triangle Only

Since correlation matrices are symmetric, showing only the lower triangle reduces redundancy:

Highlighting Strong Correlations

Correlation Significance

Practice: Financial Correlation Analysis

Key Takeaways

• Correlation matrices show pairwise relationships between all variables
• Use diverging colormaps (RdBu_r, coolwarm) centered at 0
• Show only lower triangle to reduce redundancy
• Highlight strong correlations (|r| > 0.7) for quick insights
• Add significance markers to show statistical confidence
• Correlation ≠ causation - these show relationships, not causes

In the next lesson, we'll explore clustermaps for hierarchical pattern discovery.