Home icon
Data Visualisation Guide

Choropleth classification methods

4 minutes read

Pitfalls in cartography

Choropleth maps are a popular way of visualising numerical data associated with geographical entities. Below is a choropleth map of gross domestic product per capita of EU regions.

A choropleth map of EU regions showing their GDP/capita using a pink-purple colour scheme

Source: Maarten Lambrechts, CC BY SA 4.0

This map uses 9 classes, or “buckets”, to group regions with similar gdp together. The breaks between these classes are not chosen at random: in this case they are chosen so that each class has the same interval between the class boundaries (this method of classifying is called “equal interval”). You see that in the histogram below.

A histogram showing how the class breaks in the choropleth map above are determined with an equal interval algorithm

Source: Maarten Lambrechts, CC BY SA 4.0

A better legible classification for this map is to use a “pretty breaks” classification, that uses round numbers for class breaks. Apart from the first and last classes, all classes still have the same width.

The same choropleth map as above, but with a pretty breaks classification. Each class is 10.000 €/year wide

Source: Maarten Lambrechts, CC BY SA 4.0

A histogram showing how the class breaks in the choropleth map above are determined with a pretty breaks algorithm

Source: Maarten Lambrechts, CC BY SA 4.0

By now you might have noticed that 2 regions are outliers, with a very high gdp/capita: Luxembourg and the Southern region of Ireland. Using an equal interval or pretty breaks classification might not be the best choice here, because the presence of the outliers compresses a lot of the other regions in the classes with low values.

Another classification method is to choose the class breaks in a way that every class contains the same amount of regions. This is called an “equal count” classification.

The same choropleth map as above, but with an equal count classification.

Source: Maarten Lambrechts, CC BY SA 4.0

A histogram showing how the class breaks in the choropleth map above are determined with an equal count algorithm

Source: Maarten Lambrechts, CC BY SA 4.0

The number of classes also determines the patterns visible on the map. This is the same map with an equal count classification, but with 4 instead of 9 classes.

The same choropleth map as above, but with an equal count classification with only 4 classes

Source: Maarten Lambrechts, CC BY SA 4.0

And yet another classification method is to show how far from the mean value each region’s gdp is using the standard deviation.

The same choropleth map as above, but with a standard deviation classification.

Source: Maarten Lambrechts, CC BY SA 4.0

In this case, it makes more sense to use a diverging colour palette, to show the regions with lower than average values with one colour, and regions with higher than average values with another colour.

The same choropleth map as above, but with a green - purple diverging colour scheme

Source: Maarten Lambrechts, CC BY SA 4.0

A histogram showing how the class breaks in the choropleth map above are determined with a standard deviation algorithm

Source: Maarten Lambrechts, CC BY SA 4.0

These different versions of the regional gdp map show that a choropleth map can look very different depending on the classification method and the number of classes used. The classification method should be chosen in a meaningful way that best reveals the patterns in the data.

Related pages

Normalising data for mapping

Map projections

Disputed territories

Choropleth maps

Scaled symbol maps

Cartograms

Pitfalls in cartography