Correlation & Regression Calculator

Paste two columns of numbers and see the whole relationship at once: the scatterplot, Pearson r with its p-value and confidence interval, Spearman's ρ, the least-squares line, R², and the residuals. It also checks how much r depends on any single point. Nothing leaves your browser.

Before you trust the number

• Look at the scatter, always. Anscombe's quartet (button above) shows four datasets with the same r and the same line but wildly different shapes: a curve, an outlier, a lone leverage point. The number alone can mislead. (Correlation)
• Pearson vs Spearman. Pearson r measures straight-line association and is sensitive to outliers; Spearman's ρ ranks the data first, so it captures any monotonic trend and is barely affected by extreme values. A big gap between them suggests the relationship is curved or driven by outliers.
• Check the robustness readout. If removing a single point swings r from .6 to .3, then that point, more than the relationship, is driving the result. Click it on the scatterplot to see the line and r without it, then investigate why it's there. (Regression Diagnostics)
• Correlation isn't causation, and R² isn't proof. A tidy line predicts; it doesn't explain why. Turn the relationship into a model on the Simple Linear Regression lesson, and mind the assumptions.