When a map is the wrong figure for a district comparison

Coastal municipalities like maps. The Black Sea gives a shape that looks official on a cover. Inside a survey release, that shape is often a lie. A sample of a few dozen interviews, spread across settlements that share a municipal name, cannot carry a fill colour from the mountains to the port.

We refuse a choropleth when the unit of analysis is smaller than the sample can support, when the legend uses unequal bins chosen to make one district look dark, or when the map is being used because the table felt “too dry” for a donor pack. Dry is a virtue in a sitting. A ranked table with the count of interviews in a column next to the percentage is usually the figure that can be defended.

Maps still earn a page. Administrative boundaries for a yearbook that reports every municipality in Adjara, with a full extract from a register, can be drawn as a locator — one colour, no data fill — so a reader who does not live here can see which names sit next to which. That is a map as a caption, not as a statistic.

If a mayor has already promised “a heat map of complaints”, the inventory meeting is the moment to say that the complaints file is a list of addresses with missing streets, not a grid. Replacing the promise with a ranked bar of complete districts, and a note on the streets that were dropped, has saved more than one utility unit from a second sitting on the same page.