Read a graph or map as a set of choices

A line graph can show a quantity changing with time; a scatter plot can compare paired variables; a histogram summarizes a distribution. These are different arrangements of evidence. Connecting points with a line does not necessarily mean the instrument measured continuously between them.

Read the axis labels, units and tick spacing before interpreting a slope or distance. A logarithmic axis uses multiplicative intervals rather than equal additive steps. A map also has a projection, spatial coverage and color scale that shape the visual comparison.

Suppose fictional measurements are 2, 4 and 8 units at times 0, 1 and 2 hours. The value doubles each hour. On a linear vertical axis, the increases are 2 and 4 units; on a logarithmic axis with base-two steps, both increases occupy equal vertical distances. The same data support both displays.

Now compare two maps whose darkest color represents 10 units in one and 100 in the other. Equal darkness does not establish equal values. A shared legend can aid comparison, but it must still be appropriate to each dataset’s units, valid range and purpose.

One dataset, different axis meanings

A monthly average and a daily extreme answer different questions. Averaging can make a pattern clearer while hiding short events. A map may also represent a composite assembled from observations at different times rather than one simultaneous exposure. Read the time definition before treating every pixel — open definition as contemporaneous.

A line drawn across a missing interval is an interpolation — open definition choice, not a new observation. In a fictional time series with readings at hours zero and four, the curve between them needs a stated model if used quantitatively. Preserve the distinction between recorded values, summaries and inferred values.