Data

Correlation vs Regression: When to Use Which (With Real Examples)

Correlation and regression are the foundation of every data science course — and the source of more misuse than any other pair of techniques.

Correlation: Direction and Strength

Pearson’s r ranges from −1 to +1. It tells you whether two variables move together and how tightly. It says nothing about which causes which. Use our Correlation Calculator.

Interpreting r

• |r| < 0.3 — weak
• 0.3 ≤ |r| < 0.7 — moderate
• |r| ≥ 0.7 — strong

Regression: Quantifying the Relationship

Linear regression fits the line y = a + bx through your data, letting you predict y from x. The slope b is the change in y for a 1-unit change in x. Run yours in our Regression Calculator.

When to Use Which

• Correlation — exploratory analysis, screening many variables, no prediction needed.
• Regression — prediction, quantification, controlling for multiple variables.

The Correlation ≠ Causation Reminder

Ice cream sales correlate with drownings. Both are caused by summer. Always ask: is there a confounding variable?

Worked Example

Marketing spend (x) and revenue (y) across 12 months. r = 0.84 (strong positive). Regression: y = 50,000 + 4.2x. Each $1 of marketing yields ~$4.20 in revenue — useful for budgeting, but only within the range of historical spend.

FAQs

Can I extrapolate beyond my data? Risky — relationships rarely stay linear forever.

What’s R²? The proportion of variance explained. R² = 0.71 means 71% of y’s variance is explained by x.