In the human body over 70% of our sense receptors are found in the eyes, and powerful data visualisations are a great opportunity to support and reinforce our written work, as Jessie described last week. Researchers across all disciplines spend weeks, months and even years collecting data, so it makes sense to spend longer thinking about how to display our findings.
Today I want to introduce some key principles behind data visualisation, so that tomorrow we can compare good and bad examples in more detail.
Keep it simple
Good data visualisations are accessible and clear. If charts, diagrams or graphics contain unnecessary information, they will be confusing and ultimately detract from the point you’re trying to make. Edward Tufte, a pioneer of data visualisation at Yale, argues that there should be a high “data-ink ratio”, meaning that the excessive decoration of information should be avoided. For example, consider leaving out boarders around charts or deleting unnecessary labels.
Hold back on colour
We all love a bit of colour, but consider using only one or two in your charts, and avoid garish colours where possible. Take Charles Booth’s famous poverty map of London (1889), showing the economic class of the residents: yellow indicates wealthy residents, red the middle classes, and black the ‘lowest class…occasional labourers, street sellers, loafers, criminals and semi-criminals’. Booth thought carefully about how to communicate wealth across London, and decided to display all other details in his maps (road names, parks and roads) in black and white (see image).
Use the right chart type
Picking a chart-type is often intuitive, but it’s worth exploring different options as there are certain pitfalls to avoid. Bar charts are good for comparison and clearly display data values. Line charts are useful for time-series analyses and can include trend-lines to help decipher patterns. Relationships, on the other hand, are generally displayed using scatter graphs, although these often feature trend-lines too.
Tune in tomorrow as I walk you through these different types.