What this answers

This calculator answers "which pairs among several numeric variables move together, and how strongly?" in a single table, rather than running a separate correlation calculation for every pair by hand.

How it is calculated

Every pair of columns gets its own Pearson correlation, computed using only the rows where both of that specific pair's values are present, a pairwise-complete-case policy. This means a missing value in one column only affects the pairs that include that column, not every correlation in the matrix, and different cells in the same matrix can be based on different numbers of rows, shown alongside each value.

Worked example

For three variables where x and y move together perfectly and z moves in the exact opposite direction of x, the matrix shows a correlation of 1 between x and y, negative 1 between x and z, and negative 1 between y and z, immediately surfacing the full pattern of relationships across all three variables at once.

Assumption audit

Calculated from your data: every pairwise Pearson correlation and the number of complete rows behind each specific pair.

Evidence to review: the pair count (n) behind each cell; a correlation based on very few complete pairs deserves much less confidence than one based on most of your rows, even though this table displays both the same way.

You must verify: that every relationship you care about is actually linear; Pearson correlation, computed for every pair here, can understate or completely miss a genuine nonlinear relationship between two variables.

What this result does not mean

A correlation matrix is a screening tool, not a hypothesis test; none of these correlations come with a p-value or confidence interval here, and none of them establish causation between any pair of variables regardless of how strong they look.

Common mistakes

Scanning a large matrix for the single largest correlation and treating it as the most important finding, without accounting for how many pairs were scanned, is a common way this tool's output gets over-interpreted; with enough variables, some pair will look strongly correlated by chance alone. A second mistake is ignoring the pair count shown next to each value; a correlation from only 4 or 5 complete pairs deserves far less trust than one from most of your dataset, even though both display the same way.

Limitations

This calculator reports only Pearson correlations; if your variables have a monotonic but nonlinear relationship, consider checking specific pairs with the Spearman Rank Correlation Calculator instead. Many correlations viewed at once also raise a multiple- comparisons concern this table does not correct for on its own.