Overall considerations Principles

Overview

Data visualisation refers to using visual elements, like lines, shapes and colours, to show data. It can take the form of charts or maps.

Using charts and maps can make data:

• more engaging
• easier to understand
• more memorable

This is because our brains have developed to quickly interpret and respond to visual information. Visualisations show data in perspective and make the data relatable to the user.

At the Office for National Statistics (ONS), we aim to make data visualisations that are:

• accessible to all users
• accurate
• intuitive to use
• engaging
• simple and clear to understand
• consistent

Following these guidelines will help you meet these aims. This will help users to understand the data, be better informed and increase trust and interest in your statistics.

Different types of data visualisations

Explanatory or communicative

This guidance focuses on the use of charts and maps to communicate data to the public, with the goal of sharing main trends from analysis with users. This is sometimes known as explanatory or communicative data visualisation.

Exploratory

Exploratory data visualisation is the use of charts and maps to uncover new insights in data. This can help to reveal trends in the data that are hard to uncover through other forms of analysis.

This guidance in the ONS service manual is focused on communicative rather than exploratory data visualisation. However, much of the guidance for creating effective charts in general can also help make better exploratory visualisations.

When to use data visualisation

Data visualisations often attract users’ attention. Use charts and maps to help communicate the main findings of a release and show the most important trends in the data.

Avoid using charts or maps to show information that is unimportant, or supplementary to the main findings. Do not use charts to introduce concepts or data that are not discussed in the main text of your release.

Data visualisations must work alongside the other parts of your release, such as the section headings and main text. Consider how your charts and text work together and always put charts close to the text they relate to.

The text that accompanies your chart or map must provide additional context or detail to the trend or finding being presented. Always use consistent wording in both the text of your publication and your charts. Sometimes you may need to simplify your categories or labels to ensure they are clear and readable in your data visualisation.

Read more in our chart text guidance

Showing the important comparison

Plotting values on a chart alone does not guarantee the trends will be clear to the user.

Design your chart in a way that clearly shows the important trends to the user.

Carefully consider:

• what data you show
• what chart type you use
• the axis scale
• chart text (such as titles, subtitles, and footnotes)
• chart annotations

Consider what is the most important trend or piece of information you want the person reading the chart to take away. Usually, this will be one of the main findings of your release, which should also be highlighted in the title of your chart (opens in a new tab) 

Use a comparison that helps to show this trend or finding clearly. For example, consider if it is important to compare values with:

• the situation 10 years ago
• the situation last year
• other components of a total
• other local areas or regions
• other countries
• other sectors
• the national average
• highest or lowest values

It is often easier to understand a chart when you are already familiar with the data. Your chart should make the data understandable to a non-expert user. Show your chart to someone less familiar with the data to see if they can understand the trend you are trying to show.

Keep charts simple

Aim to make your chart as simple as possible while still giving users the necessary context to understand the data.

Simpler charts tend to be easier for users to interpret. Trying to show too much in a chart can lead to charts that are over complicated and do not show the trends clearly.

Using two charts may be more effective at showing different relationships in the data, even if this means there is some overlap between charts. If the trend you are showing is not immediately clear to the user, reconsider your chart choices.

What data to include

When creating a chart or map, consider which variable or variables will most clearly visualise your main trend.

Only include the data that are needed to show the important trends clearly and accurately. Data visualisations should not be used to share your entire dataset.

Instead, include all the data in the dataset accompanying your publication.

Read more in our dataset guidance

Using derived variables

In some cases, the most effective way to convey the trends in your data is to use a derived variable. A derived variable is a variable calculated from one or more other variables. An example is percentage change, which is derived from data for two different time periods.

Other derived variables include:

• percentage point change
• difference from an average
• percentage of a total
• rate per person

When creating a visualisation, consider which variable or variables, including derived variables, will most clearly visualise your main trend.

The following example shows population values for 2011 and 2021. However, it does not clearly show the rate of change between the two periods.

This makes it difficult for users to interpret the difference between the two bars.

The following example shows the percentage change in the population between 2011 and 2021 rather than data values for each period:

By presenting the percentage change rather than data values, and ordering the percentage change from highest to lowest, users can clearly see how fast the East Midlands has grown compared with other regions.

Read more about the importance of ordering our chart ordering guidance

Help improve this page

Let us know how we could improve this page, or share your user research findings. Discuss this page on GitHub (opens in a new tab)