What Is a Choropleth Map?

A choropleth map colours predefined geographic areas according to a numeric value. Countries, districts, municipalities or census tracts are shaded with lighter or darker colours so readers can compare a statistic across those areas.

The method works best when the mapped values are genuinely comparable between polygons: rates, percentages, densities, ratios or indexes. A choropleth is less suitable for raw totals because polygon size, population and exposure can dominate the pattern.

How the map works

Each polygon receives a value, which is then mapped to a colour scale either continuously or through a small number of classes.

For example, a map of vaccination coverage might colour districts from light to dark according to the percentage of eligible residents vaccinated. The polygon represents the district; the colour represents the district-level statistic.

The map does not imply that every location inside the polygon has the same underlying reality. It says that the reported or calculated value applies to the area as an aggregate.

What kinds of values fit choropleths?

Good candidates include:

percentages, such as the share of households with electricity;
rates, such as incidents per 100,000 residents;
densities, such as population per square kilometre;
ratios, such as doctors per 10,000 people;
indexes that have been constructed for comparison across areas.

Raw counts require more care because a district with twice the population can have twice as many events even if the underlying risk is identical. Why Raw Counts Usually Should Not Be Mapped with a Choropleth explains that issue in detail.

Classification changes what readers see

Many choropleths group values into colour classes, and where those class boundaries fall affects which regions appear similar and where the map creates visual contrast.

Quantiles, equal intervals and natural breaks answer different classification needs. An unclassed choropleth can instead use a continuous colour ramp.

Those decisions belong to the classification method rather than the definition of the map itself, but they are not cosmetic. Natural Breaks, Quantiles, and Equal Intervals covers the trade-offs.

Boundaries are part of the analysis

A choropleth depends on a system of zones, so changing those zones can alter the visible pattern even when the underlying observations do not.

A neighbourhood-level rate may reveal variation hidden at district level, while different district boundaries can aggregate the same observations differently. This is part of the modifiable areal unit problem (MAUP).

The map therefore communicates two things at once: the statistic and the geographic partition used to summarise it.

What a choropleth should not imply

A choropleth should not make readers believe that:

the value is uniform inside every polygon;
a boundary marks a sudden real-world discontinuity;
a dark large polygon necessarily contains more total events than a small one;
missing data are the same thing as zero;
the chosen class breaks are natural facts rather than analytical choices.

Legends, notes and careful variable selection help prevent those interpretations.

Choropleth vs other thematic maps

Use proportional symbols when the main variable is a total attached to places, a heat map when the underlying observations are points and concentration matters, and a dot-density map when you want to communicate quantity distributed within areas without implying exact individual locations.

The choropleth is specifically an area-comparison map.

A simple decision rule

Use a choropleth when:

the data are meaningfully attached to predefined areas;
the value is comparable across those areas;
the reader should compare area-level patterns;
the boundaries themselves are an acceptable frame for the question.

If the statistic only makes sense as a total, or if the phenomenon is fundamentally point-based, another map type is often clearer.

Classification and colour are separate decisions

A choropleth needs both a statistical mapping rule and a visual palette: the classification determines which values are grouped together, while the palette determines how those groups or continuous values are perceived. A poor classification cannot be repaired by better colours, and a sensible classification can still become difficult to read if the palette does not preserve order.

For a sequential statistic, darker or otherwise perceptually stronger colours should normally correspond consistently to higher values. If the variable has a meaningful central reference such as zero change, a diverging scheme may be appropriate instead.

Choropleths are sensitive to the denominator and the boundary version

A rate can be mathematically correct but still be difficult to compare if numerator and denominator refer to different dates or boundary systems. If cases use 2025 districts but population uses a 2020 district table whose geography changed, the apparent rate can be a join artefact rather than a real geographic pattern.

Before styling, confirm that the statistic, denominator and polygon geography refer to compatible units and periods. This is especially important when administrative boundaries have been split, merged or renamed.