Skip to content
MangoMap Help Centre
Esc
↑↓navigate↵open⌘Jpreview
On this page

Choosing a classification method for choropleth maps

Choose between Equal Interval, Quantile, Natural Breaks, and Head/Tail Breaks when you colour points, lines, or polygons by a number.

Equal Interval, Quantile, Natural Breaks (Jenks), and the head/tail method all give valid maps. They also give visibly different maps from the same data. Choose the method that fits your data and the question the map answers.

See it on an example map

Open the World Classification Lab. Population — Equal Interval is on when the map opens. Turn it off and turn on Population — Quantile. Do the same with the other layers in Population Classes, and see how the same countries change colour.

The classification method works the same way on every geometry: countries coloured by population, roads coloured by length, or cities coloured by population.

This page explains how to choose a method. The companion reference, Classification methods for by-quantity styling, defines each method.

The same data, four maps

Take a polygon layer of 177 world countries with a pop_est population field. The smallest value is 140. The largest is about 1.4 billion. Half the countries have fewer than about 10 million people, and a few have hundreds of millions.

Colour the layer by pop_est with 5 classes. Here’s what each method does to the same data:

  • Equal Interval splits the range into 5 bands of about 280 million each. 174 countries fall in the first band. Two middle bands are empty. Only China, India, and the United States stand out. The map is true to the absolute numbers, but it hides every difference between typical countries.
  • Quantile sorts the values and puts about the same number of them in each class, here about 35 countries. Every colour covers a similar share of the countries, so the map shows rank: the bottom fifth, the middle, the top fifth. Class boundaries fall at uneven numbers. The map gains comparison; the legend loses round figures.
  • Natural Breaks (Jenks) finds gaps in the distribution and puts the class breaks there. The map shows the groups the data falls into naturally.
  • Head/Tail Breaks makes classes that get wider from the lowest class to the highest. The lowest class is narrower than with Equal Interval, but on this data 162 countries still fall in it. The map gives a little more detail among the smaller values, and the largest countries still stand alone.

The same numbers give four maps. Each method answers a different question.

To see why the maps are different, compare where each method puts its class breaks and how many countries each class gets.

Same countries, four classification methodsThe populations of 177 countries on a log scale, cut into 5 classes four ways. Countries per class, lowest class first: Equal Interval 174, 1, 0, 0, 2; Quantile 36, 35, 36, 35, 35; Natural Breaks (Jenks) 36, 35, 53, 51, 2; Head/Tail Breaks 162, 12, 1, 0, 2.10010,0001 million100 million1 billionPOPULATION, LOG SCALE: EACH GRIDLINE IS 10 TIMES THE ONE BEFORE177 COUNTRIES, ONE DOT EACHCHINA, INDIACOUNTRIES IN EACH CLASS, LOWEST CLASS FIRSTEqual Interval1741002Quantile3635363535Natural Breaks (Jenks)363553512Head/Tail Breaks16212102

Each tab shows the same layer with one method, as a viewer sees it, with its legend.

World Classification Lab with Population — Equal Interval shown: almost every country is the palest yellow and only China and India are dark red; the legend lists the classes >=140, >279543112, >559086084, >838629056, and >1118172028
World Classification Lab with Population — Quantile shown: countries spread across all five colours from pale yellow to dark red; the legend lists the classes >=140, >2494530, >6855713, >16296364, and >39309783
World Classification Lab with Population — Natural Breaks shown: many countries are orange or red and China and India are dark red; the legend lists the classes >=140, >2494530, >6855713, >25666161, and >328239523
World Classification Lab with Population — Head/Tails shown: most countries are pale yellow, a few large countries such as Russia, Nigeria, and Indonesia are light orange, and China and India are dark red; the legend lists the classes >=140, >93181130.66666667, >279543112, >559086084, and >931810046.6666667

Match the method to your data’s shape

  • Values spread evenly across the range. Equal Interval works, because equal-width classes hold real slices of the data. Rainfall measured at weather stations (points) can look like this.
  • A few large values and many small ones. Equal Interval puts most features in one colour. Country population (polygons) looks like this. So does road length (lines): half the Singapore road segments are under about 30 metres, and a few run for several kilometres. Pick Quantile (every class holds features) or Natural Breaks (Jenks) (class breaks at gaps) instead.
  • Clear groups in the data. Natural Breaks (Jenks) is built for this. It finds the gaps and puts the class breaks there.
  • You want simple ranges with more detail at the low end. Head/Tail Breaks keeps ranges tied to the number line, like Equal Interval, but makes the low classes narrower.
  • You don’t know the shape yet. Try Natural Breaks (Jenks) first. It is the default when you pick a numeric field, and it gives sensible class breaks for most real data.

Match the method to your reader’s question

The method also sets what the map shows. Four viewer questions, four method picks:

  • “How does this feature compare to the others?” Quantile. Each class holds a similar number of values, so the colour shows rank.
  • “Where are the natural groupings?” Natural Breaks (Jenks). The class breaks fall at the gaps.
  • “How large is this value on an absolute scale?” Equal Interval. Use it when the width of a class means something: rainfall in mm, temperature in degrees, elevation in metres.
  • “How large is this value, with more detail among the small ones?” Head/Tail Breaks. The classes stay tied to the number line but start narrow.

To pick fast, answer these questions in order. The first yes gives your method, and Natural Breaks (Jenks) is the answer when none fits.

Which classification method to pickFlowchart: pick Quantile when the colour should show rank, Equal Interval when the values spread evenly across the range, Head/Tail Breaks when you want more detail among the small values, and Natural Breaks (Jenks), the default, for clear groups or when you are not sure.YESYESYESNONONO OR NOT SUREColour by a number in classesShould the colourshow rank?Values spread evenlyacross the range?More detail amongthe small values?Quantileabout the same number of valuesin every classEqual Intervalclasses of equal width, such asrainfall in mm or height in metresHead/Tail Breaksclasses widen from low to high(Head/Tails on point layers)Natural Breaks (Jenks)the default: class breaks at the gaps,for clear groups or when unsure

When in doubt, try two and compare

The quickest way to choose is to compare the results. Style the layer with Natural Breaks (Jenks), save, and look. Style it with Quantile, save, and look. The right one is usually clear once you see both. The wrong one puts most features in one colour or makes the legend hard to read.

Switching the classification method rebuilds the class breaks from scratch. Any changes you made to a class break’s colour, range, or label are lost. Compare methods first, then customise.

What this isn’t

This page doesn’t replace Classification methods for by-quantity styling. That page defines each method and how it calculates the class breaks. This page helps you choose between them.

It also doesn’t cover How to customise category and class break symbology. Once you’ve picked a method, you can round awkward boundaries (17.342 → 20), rename labels, or change colours. That’s a separate task.