The chart that made categories visible

A chart can argue before its reader has interpreted a single number. Florence Nightingale’s 1858 mortality diagram is a good example because its design makes the categories and the comparison rules part of the evidence.

The plate contains two circular diagrams. The right circle covers April 1854 to March 1855. The left covers April 1855 to March 1856. Each month occupies a wedge, and the categories are drawn from the same centre. The legend says that the blue, red, and black wedges represent deaths from the printed categories of preventable or mitigable zymotic diseases, wounds, and all other causes. Those historical terms belong to the artifact and should not be silently modernized.

The legend also gives the reader a method. The blue, red, and black wedges are compared area for area, with the centre as their common vertex. The chart does not mean that a longer radius alone represents proportionally more deaths. It asks you to compare the areas enclosed by the coloured boundaries. Nor is it asking you to add three stacked slices into one monthly bar. It asks you to follow the blue, red, and black lines around the circles. In two cases the legend calls out coincidences: in October 1854 and April 1855, the black area coincides with the red; in January and February 1855, the blue coincides with the black. The instructions are part of the chart’s argument because they tell you how to see categories that occupy the same radial extent.

I find the design unusually candid. It does not pretend that a category label is self-explanatory, or that a circle automatically teaches its own measurement. The legend tells you what each colour means, how the areas are measured, and how to compare the boundaries. A reader who skips the legend can still see a pattern, but cannot be sure what the pattern represents.

The plate does not, by itself, prove why mortality changed, and I have not reconstructed its values or checked the underlying population denominators. It shows a particular classification and a particular visual comparison across two periods. That is enough to teach a useful habit: before asking what a chart says about the world, ask what its geometry says you are allowed to compare.

A modern chart can have cleaner colours, tooltips, and a legend that appears on hover. The obligation is the same. State the categories. State the measurement. State how overlaps should be read. A chart becomes persuasive through those choices, and the reader deserves to see them.