How to avoid being deceived by the ways charts can lie
To finalise our series on values, I offer a book review to help analysts not lie. Well, actually it’s as much to help leaders avoid being deceived.
The book in question is the brilliantly timely “How Charts Lie” by Alberto Cairo. Those familiar with the field of Data Visualisation will already know Alberto’s name from either his presentations or his past books “The Truthful Art” & “The Functional Art”.
Alberto is the Knight Chair of Visual Journalism at the University of Miami. How cool a job title is that? With a background in journalism, in this book, he turns his attention to the layperson.
During a time when politicians & social media have been seen to share deceptive statistics and data visualisations, this book is a welcome defence. In this very accessible short book (under 200 pages excluding appendices) any leader can learn to have a better critical eye when presented with charts.
We need to improve our Graphicacy
One would be surprised to meet a senior commercial leader who struggled with their numeracy or literacy. Yet we still accept leaders who can’t really grasp what charts & other graphics mean, nor how to critique them.
Through an introduction packed with timely political examples (albeit all US-based), Alberto makes clear the scale of the problem. From all sides of political debate, he shares examples of charts that lie. Ones that misrepresent the actual data or statistics in different ways.
This book is a plea for all citizens to take the skill of Graphicacy seriously. We live in a society awash with so much data that we rely on analysis & data visualisations to help us make sense of it all. Graphicacy competency is essential if we are not to all be at the mercy of the visualiser.
In printed form, there is also a beautiful simplicity to the way all his charts are shown in only black, white & red (plus the use of saturation to add additional information). It makes for a clear & practical book that walks the reader through all types of basic charts, maps & infographics, many of which will look familiar.
Understanding the component parts & how to read
As someone who trains others in Data Visualisation, I was very impressed by how many basics Alberto covers in his first chapter. Entitled ‘How Charts Work‘, it walks the reader through how to understand and read a graph, map or data visualisation.
Covering a little history of data visualisation, he goes on to explain the component parts:
- Scaffolding (titles, legend, scales, bylines)
- Content (including use of Visual Encoding)
- Annotation layer
Alberto then uses awareness of the broad options for visual encoding data to help the reader read a wide diversity of charts, From scatter plots to bubble charts. From heatmaps to the dreaded pie chart. From line charts to the more advanced connected scatter plot.
Edward Tufte would be pleased to also see the coverage of the power of small multiples. Alberto also goes on to share a great 5 step approach to reading charts. In order, you should look at the…
- Title, introduction (or caption) and sources
- Measurements, units, scales & legends
- Methods of visual encoding
- Annotations
- A birds-eye view of the whole (to spot patterns/trends/relationships)
Watch out for these ways that charts may lie to you
In the remaining chapters, Alberto brilliantly guides the reader through a number of the different ways that charts can lie. I won’t steal his thunder by sharing too much with you (as I want you to read his book). But he does this without prejudging people’s motives & with a balance that does not favour any political camp.
His chapters explore the following ways that charts can lie to readers:
- Charts that lie by being poorly designed
- Charts that lie by displaying dubious data
- Charts that lie by displaying insufficient data
- Charts that lie by concealing or confusing uncertainty
- Charts that lie by suggesting misleading patterns
In all 5 chapters, he shares plenty of examples to demonstrate both how this happens & what it can look like in practice. He is honest about his own mistakes, while also sharing a passion for us to all do better.
I’m sure when analysts read this they will pick up the sense of moral obligation to avoid such mistakes & keep their charts as honest as possible. But, Alberto’s very practical book can help not only those designing data visualisations but also those reading or crucially sharing them. His plea for more care before sharing on social media was a conscience pricker for me.
How can you help join the Graphicacy Revolution?
In conclusion, Alberto shares the story of someone who may be a surprising Data Viz heroine to some – Florence Nightingale. Reproducing her famous “Causes of mortality in the armies in the east” chart from the Crimean War, Alberto explains its design.
However, it is his key message is not so much the chart itself, but the ethical example of Nightingale. Faced with the reality of how many lives have been lost due to poor sanitation, she takes the very approach Alberto advocates throughout this book. She checks the data & statistics thoroughly & then clearly shares the truth, even if that makes her look bad.
How can you too become more serious about your responsibility when designing or sharing charts. If you want to improve your graphicacy & protect yourself against being misled by charts – I recommend you read this book. It might just prove to be a public service.
Finally, I should also mention that it’s an enjoyable and immersive read. A book that I read in under 2 weeks without much of an effort.