Misleading maps rarely look wrong, which is why you can trust them too easily. Check whether values are rates, how the classes are cut and how many colors are used, and most distortions show up.
This guide sits within the wider topic of geospatial data visualization, and it focuses on the design choices that bend a map’s message.
| Item | Value |
|---|---|
| Typical bands on a choropleth map | 5 color bands |
| Maximum colors in a diverging scale | up to 6 colors |
| Safe range of classes | 3-7 data classes |
| Standard break method in the ONS manual | Jenks natural breaks |
These figures come from the ONS service manual and the Axis Maps cartography guide. Guidance can change, so check the sources before you rely on them.
Why maps mislead without anyone meaning to
A map feels objective because it looks like the ground itself. In practice, every thematic map rests on a stack of design choices: what to count, how to group values and which colors to use.
The Axis Maps guide is blunt about the stakes: maps that use questionable classification methods are more than ineffective, they are misleading. It adds that readers rarely question the classes on a map, which makes classification one of the easiest ways to lie with maps, whether intentionally or accidentally.
The sections below cover the places where distortion usually hides:
- the measure being mapped, such as totals instead of rates;
- the way values are grouped into classes;
- the number and type of colors;
- the decision to use a map at all.
Raw counts versus rates
The most common problem is mapping totals. The ONS service manual says choropleth maps should show standardized rates rather than absolute values, because absolute values are distorted by the size of the area.
Its example is unemployment. A rate gives the percentage of the population unemployed in each area, while a total simply tracks where more people live.
In most cases the ONS standardizes by population, though land-use data can instead use a rate per square kilometre. A map of totals mostly redraws the population map, so the biggest or most crowded areas look like the biggest problems.
You can compare the main map styles in this overview of types of thematic maps.
How classification changes the picture
Classification groups many observations into a few ranges. According to Axis Maps, small differences in how you do this can dramatically change the look of the map, and therefore its message.
Each common method has a typical failure:
- Equal interval splits the range into equal-size classes and works best on data spread across the whole range. It is a poor choice for skewed data or one or two large outliers, which leave empty classes.
- Quantiles put the same number of places in every class. The catch is that classes can have very different numerical ranges, such as 1-4, 4-9 and 9-250, and similar places can land in different classes.
- Natural breaks minimize variation within classes and maximize differences between them. Each dataset gets its own solution, which makes comparison across a series of maps awkward.
- Manual breaks let you hardwire important thresholds, or keep the same classes across a series so that colors always refer to the same numbers.
The ONS uses Jenks natural breaks as its standard, noting that it works well for most datasets. It also allows other methods or fixed values, for example to separate areas above and below zero (ONS).
Before you pick a method, look at a histogram or scatterplot of the values. Axis Maps recommends this to see the form of the observations, and gaps in the histogram suggest natural places to break the data (source).

Colors, bands and the limits of the eye
The ONS typically uses 5 color bands for choropleth maps, and a diverging scale can include up to 6 colors. It advises against a fully continuous scale or too many bands, because readers then struggle to tell values apart (ONS).
The same limit appears elsewhere. Axis Maps suggests 3-7 data classes to be safe, and Esri’s ArcGIS Pro documentation calls 7 colors the approximate upper limit that can be clearly distinguished on a map. Esri adds that too many classes cause trouble, especially with light colors.
Too few classes distort in the opposite direction. Axis Maps warns that very coarse schemes can create artificial geographic patterns by lumping together places that are quite different (source).
Match the palette to the data. A sequential palette runs from low to high, while a diverging palette uses two contrasting colors around a neutral midpoint (ONS).
Diverging palettes matter when the data pass through a meaningful value. Esri notes that data showing positive and negative change may vary around a value of 0, and a critical break there makes the turning point obvious (Esri).
Is a map the right chart at all?
Having geographic data does not automatically make a map the best choice. The ONS manual suggests a map when the pattern is primarily geographic, such as north versus south or urban versus rural. With little geographic pattern, the map may look noisy and fail to help users.
The manual also advises against maps when you have data for only a few areas, such as the four nations of the UK or the nine regions of England. A bar chart makes those comparisons easier.
For change over time, the ONS suggests calculating the rate of change and showing it on a single map instead of two separate maps. If a map does not show clear patterns, it proposes a scatter plot for relationships or a bee swarm chart for distributions.
If your map will sit beside filters and other charts, read about geospatial dashboard design next.
A quick checklist for reading any map
Run through these steps before you accept what a map seems to say:
- Find out whether the measure is a rate or a total, and what it was divided by.
- Read the legend: how many classes there are and where the breaks fall.
- Ask which classification method was used and whether it suits the shape of the data.
- Check whether the palette is sequential or diverging, and whether any midpoint means something.
- If several maps are compared, confirm that they share the same classes.
- Ask whether a bar chart or scatter plot would show the point more honestly.
A map that fails several of these checks deserves to be treated as a hypothesis rather than a finding.
Frequently asked questions
What makes a map misleading?
Show answer
Usually the data are fine and the design is the problem. Typical culprits are totals shown where rates belong, class breaks that favor one story, an unhelpful number of colors, or a map used where a chart would work better.
How many classes should a choropleth map have?
Show answer
Axis Maps treats 3-7 classes as a safe zone, and the ONS default sits inside it. Look at a histogram of your values to decide where within that range to land (Axis Maps).
Should I map totals or rates?
Show answer
Choose rates. Totals follow the size of each area and its population, so they say little about how intense a problem is (ONS).
Does the classification method matter if the data are accurate?
Show answer
Yes, because the same accurate numbers can tell different stories under different breaks. Label the method in the legend, and reuse identical breaks when maps are meant to be compared (Axis Maps).
When should I use a chart instead of a map?
Show answer
Switch when you have only a handful of regions or when the data show no real geographic pattern. In those cases a bar chart or scatter plot is easier to read (ONS).



