kpimaster
Open KPI Master

Are these two columns actually related?

A scatter plot, a line and one number answer the question better than a hunch. This guide explains what the number means, when it misleads, and where to find it in KPI Master.

Open the working example

The example uses illustrative order data bundled with the tool. It is not a customer result.

1. Correlation measures co-movement, not cause.

r summarizes how closely paired values follow a straight line, from −1 (one rises as the other falls) through 0 (no straight-line pattern) to +1 (they rise together). KPI Master puts a word beside it: weak below 0.5, moderate from 0.5, strong from 0.7 and very strong from 0.9, in either direction.

Moving together never proves that one thing causes the other. Both may follow a third factor, or the link may run the other way. The tool words its findings as “move together” for exactly that reason.

2. Look at the dots, not only the number.

Each dot is one row. A tight band along the line is a strong link; a cloud is a weak one. Beside the chart you get r, the strength word, and the number of points that had both values. Rows missing either value are left out of that pair and never treated as zero.

When the dots form a curve or a U, a straight-line number understates the link. If a column pair looks related by a different measure (mutual information) but r is low, the tool adds a note: “curved relationship”.

3. Check these before you repeat a finding.

4. Open the Connections view.

After a file opens, choose Connections. The map shows which measures move together; tap one to see its ranked links. Compare any two columns lets you pick a pair yourself and chooses the right picture for the types: a scatter for two numbers, bars for a category against a number, a grid for two categories. The grid of all links at a glance sits below.

Everything is computed in your browser; the file is not uploaded.