How Outliers Affect Choropleth Classification

An outlier can dominate a choropleth by stretching the numeric range far beyond the values of most areas. The result may be a map in which one region receives the darkest colour while nearly every other region looks similar.

That can be an accurate reflection of an extreme value, or it can hide the variation the map was supposed to reveal.

Equal intervals are especially sensitive to range

If most values lie between 10 and 30 but one value is 150, equal intervals based on the full range allocate much of the colour scale to values where almost no observations exist. As a result, most areas can collapse into one or two low classes.

This does not mean equal intervals are wrong; it means the range is being controlled by the extreme value.

Quantiles keep colours populated

Quantile classification will still place roughly equal numbers of areas in each class, so the map remains visually varied. However, the top class can contain values that are numerically very different: a value of 35 and the outlier of 150 may share a colour simply because both fall within the highest-ranked group.

Quantiles solve the empty-class problem, not the outlier's substantive interpretation.

Natural breaks may isolate the extreme

Natural-breaks methods often create a separate class around an extreme value if the gap is large enough. That can be useful because the outlier becomes visibly exceptional while the remaining classes describe the main distribution more clearly.

But a one-feature class deserves explanation. The reader should know whether the extreme is meaningful, an error or a consequence of a very different denominator.

Diagnose before redesigning

Before changing the map, inspect the outlier itself.

Ask:

Is the value valid?
Is the unit consistent?
Is the denominator unusually small?
Is the area genuinely exceptional?
Does the outlier belong to a different process or geographic category?

Never suppress an extreme merely because it makes the legend inconvenient.

Possible responses

Depending on the question, you can:

use natural breaks to isolate the outlier;
use a transformation such as logarithmic scaling where substantively appropriate;
annotate the extreme value explicitly;
show a companion chart or table;
map the main distribution separately while clearly retaining the outlier in another view;
use an unclassed map with a carefully chosen scale.

The response should make the distribution easier to understand without pretending the extreme value does not exist.

An outlier can be geographically important even if it is statistically awkward

If one coastal district has an exceptionally high value because a major port is located there, isolating it visually may be exactly what the map should do. Treating it as a nuisance and clipping it into the ordinary range would hide a real geographic mechanism.

The design question is therefore not “how do I stop the outlier ruining the colours?” but “is the extreme value valid, and how should its exceptional status be communicated?”

Consider a companion distribution view

A histogram or box plot beside the map can show whether the extreme is isolated and how the rest of the values are distributed, often more clearly than forcing the legend to explain the full statistical shape.

Maps are good at showing where the outlier occurs; charts are often better at showing how extreme it is numerically.

Do not confuse outliers with data errors

An extreme value deserves verification, but unusual geography often produces legitimate extremes: small island jurisdictions, capital districts, mining regions or specialised industrial zones can differ sharply from surrounding areas for real reasons.

Check units, joins, denominators and source records first. Only after the value is validated should you decide whether the map needs a special class, annotation or alternative scale.