Presenting data visually for a poster or presentation

Visualising data and tables can enable patterns, trends, outliers and correlations to be revealed, and comparisons to be made more easily. William Playfair’s first attempts at bringing line, bar and pie charts to mass audiences in the ‘Commercial and Political Atlas’ were published in 1786^[1]
.

Since then, data visualisation has become an important aspect of research and academia. The ability to understand science often comes down to interpreting the data and following the story — and so storytelling within science communication is paramount. This is enhanced by the correct choice of data and type of graphic used, which is important not only to show the research, methods and results but also to engage readers across the wide range of media now available, including articles (both print and online), presentations, posters and social media.

If presented well, visuals are highly effective in conveying information and telling a story. If they are not presented well, they can misinform or hide the data, therefore confusing and turning the reader away.

The human brain starts processing visual cues within a tenth of a second^[2]
so often an individual will have questions about an image before they have begun to pay conscious attention to it. These questions may include: Am I interested in this? How long will it take me to read? Are the visuals interesting? Do I like the colours used? Do I understand it? Do I want to understand it? Is it telling me a story that I can follow and am I willing to invest the time in looking at this?

With this in mind, the fundamentals of producing a poster or a presentation do not differ from one another in basic principles. Both must initially attract attention but then also ensure that they have a narrative and show the results and conclusions, while keeping the reader interested. It may be helpful to try to write one long sentence describing the gist of the story as a guide. The natural breaks can become the working titles for each image — this can be an effective way of planning a visual story. Knowing who your audience is or gauging who it might be, including its familiarity with the subject matter, can help in this process.

A poster is likely to be presented in a dynamic setting, where the audience may only spend a couple of seconds looking at it and deciding whether to invest more time in reading it and absorbing the information. The poster therefore needs to outline clearly what it is about using explanatory headlines and arresting and informative visuals.

A presentation is likely to be explained by the presenter as it progresses, so viewers can be guided through the slides. This means that information can be broken up on to several slides to explain the story or process better, because additional slides come at no extra cost. Five or six separate slides explaining a process or result is far better than one slide with everything on it.

Importance of data visualisation

This basic number grid makes it difficult to see how many number 3s there are.

However, just adding colour and editing makes this clear. This is a simple but effective way to show how colour and chart type (although this example is not a chart) can enhance data and tell a story.

Probably the most famous example of this is Anscombe’s quartet^[3]
. This comprises four datasets that have near identical statistical properties. Constructed in 1973 by Francis Anscombe, it demonstrates the importance of graphing data before analysing it, as well as the effect of outliers on statistical properties.

This is still used today to illustrate the importance of looking at a set of data graphically before analysis.

Editing the data and the graphic

The editing process is vital and can make or break the visual understanding. There may be hundreds or even thousands of data but it may not be necessary to show them all. It is only important to show the whole data set if it helps the viewer to understand the story. Otherwise, it may be preferable to highlight a selection of data or even just one result to demonstrate a point. For example, line graphs that include every line and data point could be redesigned by using an aggregated line or by highlighting the most important lines as focal points, then greying out the remaining lines. This is dependent on the data and the format of the image used — it is important not to cherry pick the data or mislead the audience.

It is also important to consider the type of graphic used — for example, is a bar chart, line graph, histogram or scatterplot sufficient, or is a more complex graphical representation required? Bar graphs can be useful for comparing values, slope graphs are better for analysing trends, and column or line graphs can be used for either. Stacked bar or column graphs work well for examining the composition of data — a pie or donut chart can also be used for this as long as there are no more than a few segments. Scatter plots, bar charts, bubble plots and heatmaps are useful for investigating relationships or the distribution of data.

There are many ways of displaying the same data, some better than others, so it is important to consider what the chart is trying to show and which format works best for that. Additionally, a more general audience may be familiar with reading simple charts whereas experts might be used to more complex types of visualisations.

The colour palette used can also make a big difference. The correct palette can help guide viewers through the story and organise the graphics, reinforce the topic and provide clarity, insight and context. Colour can grab attention and can signal where to look but it should be used sparingly so it does not confuse the audience.

Colour can also be used in a quantitative way and can carry tone and meaning. For example, using one colour and tints of that colour sequentially, accenting the highest value as a dark shade and the lowest value as a light shade (or vice versa), can portray a range of data effectively. Using an incorrect palette can hide or mislead, create a barrier to understanding or even convey the wrong message — for example, some of the rainbow colour scales that are frequently used on maps do not show the data in the best way.

It is sensible to start by using only one colour and tints to see if it will work. Black and greys are a good starting point.

Regardless of who the audience is, the visuals need to describe and explain what is going on without requiring large amounts of explanatory text. A succinct explanatory heading and subheading should be sufficient.