Counts versus rates

The same numbers, mapped two ways. On the left, raw counts: the map is bright where the people are, which is nearly always what a count map shows and almost never what its author meant. On the right, the same counts divided by population, which is the actual story. This is the most common way to publish a misleading choropleth, so normalizeBy is one option, and the legend retitles itself when it is used so the two maps cannot be confused for each other.

…

The dev-mode nudge

A choropleth whose values are large integers with no normalisation gets a note in the dev-mode diagnostics, because that shape of data is a count nine times out of ten. It is advice, not a block: sometimes a count map is exactly right, and the library is not in a position to know. The bubble series is the other answer, since a circle's area is immune to how big the country under it happens to be.

series: [{
  name: 'Cases',
  joinBy: ['iso_a3', 'code'],
  data,                          // [{ code, value, population }]
  normalizeBy: 'population',     // legend becomes "Cases (per population)"
}]