Correlation Explorer

A diverging heatmap of every numeric pair in a dataset, with a click-through scatter so you can check whether the number is telling the truth.

Loading…

About this tool

Correlation matrices are the fastest way to get a wrong idea about a dataset, because a single number cannot tell you about a curve, a cluster, or one influential point. So this one makes you click through: every cell opens the underlying scatter, and Pearson sits next to Spearman so a monotone-but-curved relationship is obvious at a glance.

Worth knowing

  • A big gap between Pearson and Spearman means the relationship is monotone but not linear.
  • The palette is blue↔red through a neutral grey, so zero reads as nothing rather than as a colour.
  • Correlation on a dataset with subgroups is a trap. Try colouring by a category and see it break.