Last week, I said that knowing a bit about how human perception works can help you to make your data presentation more effective. This week, I want to consider that process in practical terms, with the help of one of the great thinkers, writers and practitioners in the dark arts of data visualisation, Stephen Few.  Few has written a useful article on ‘Data Visualisation for Human Perception’ for the online encyclopedia of interaction design, where he poses a question that anyone designing a chart, graph or figure should ask themselves – “Why am I doing this?”

Now I’m not asking you to resort to existential philosophy (drop me a line if you’d like to chat existentialism), but Few distinguishes between the visualisation work you do in order to understand or analyse your information, and the way that you present and communicate that information to others.  You often use the same tools and skills to explore data as you do to present it, but the aim, and therefore the approach, differs, so your first answer to Few’s question should be "I'm exploring my data!" or "I'm communicating this information to others!".  

If you’ve established that you are designing a figure to communicate with others, you need to clarify exactly what it is you wish to communicate, and to keep that aim in mind.  It’s important to spend some time thinking about this, particularly if you’re familiar with your data, and especially if you’re the expert on it.  Few also talks about a step in the design process which is often rushed or overlooked, where you check that what you’ve designed actually meets your aim, either on your own assessment, or by checking with a colleague, friend or family member.  Few’s example is to try and communicate the primary causes of death for Americans in a particular year.  He sets out a five point checklist that his figure needs to meet:

1. Clearly indicates how the values relate to one another, which in this case is a part-to-whole relationship - the number of deaths per cause, when summed, equal all deaths during the year.
2. Represents the quantities accurately.
3. Makes it easy to compare the quantities.
4. Makes it easy to see the ranked order of values, such as from the leading cause of death to the least.
5. Makes obvious how people should use the information - what they should use it to accomplish - and encourages them to do this.

He uses a popular figure – a pie chart – and checks against his list.  His assessment is that it fails on three of his five criteria, so back he goes to the drawing board, and tries again, this time with a bar chart.  This time, it’s five out of five, so it’s a success.  He provides a detailed analysis of why his first figure failed, and the second succeeded, and it’s worth having a read of the kinds of things you might want to consider when you’re reviewing your own visualisations.  Few's article also considers some of the tips and tricks for using our understanding of human perception, such as Gestalt, which we discussed last week.  

Few also reminds us that there may be occasions when using a table of numbers is what’s needed to communicate your message, for times when you want to communicate precise numbers, or for completeness.  But, as Warren discussed in his post last week, tables often lack the ability to get across the bigger message, the trends, the outliers.  So when you’re designing a chart or graph, the final answer to the question “Why am I doing this?” should be “Because a table just won’t do”.