A relationship chart can look plausible while its axes describe the wrong variables. Before styling it, write down what one point represents and which quantity belongs on each axis. For a bubble chart, define the third quantity that controls size as well.

Use a small practice table with three numeric columns labelled X position, Y position and Quantity. Give the rows unmistakable values: X values 1, 3 and 10; Y values 2, 8 and 4; quantities 5, 10 and 20. These are invented plotting checks, not measurements from a real study.

In Numbers on Mac, add a scatter or bubble chart from Chart, then Add Chart Data. For one scatter series, the first selected column supplies X and the second Y. For one bubble series, the third selected column supplies bubble size. Select complete corresponding records and click Done.

For multiple series, inspect Share X Values in the data-reference controls. A shared X column is appropriate only when the series really use those same coordinates. Independent coordinates require the corresponding X data for each series. Review the setting rather than inferring it from the chart’s appearance.

Check identifiable points before interpreting the pattern

Locate the point at X=10, Y=4. It should be farther right than X=3, Y=8, but lower on the Y scale. If these relationships are reversed, inspect the selected columns and axis labels before drawing any conclusion from the pattern.

For the bubble version, confirm that the row with quantity 20 produces a larger bubble than the row with quantity 5. Do not interpret bubble diameter by eye as a precise numeric ratio; the source values remain the basis for exact comparisons.

Change one practice X value and observe which point moves horizontally. Restore it, then change one Y value and check the vertical change. These controlled edits help distinguish a coordinate-reference problem from an unexpected but valid dataset.

Keep each dimension understandable

Replace the practice data with the intended records only after the mapping is clear. Label the axes and explain the size variable with its units. Inspect an extreme value and a central value against the source table. A visually convincing cluster does not establish causation, and correct formatting cannot rescue dimensions that were assigned to the wrong columns.

Reference: Apple: scatter and bubble charts in Numbers.