Instead of just looking at charts, as if they were mere illustrations, we must learn to read them and interpret them correctly. Alberto Cairo
How Charts Lie: Getting Smarter about Visual Information is about becoming an educated reader and a critical consumer of charts. This is important because, as Cairo says, “Charts – even those not designed with ill intent – can mislead us.”
Cairo, a journalist and designer, is a leading communicator in the world of data visualization. He teaches at the School of Communication at the University of Miami and has authored two popular data visualization volumes, The Functional Art and The Truthful Art. He has taught Massively Open Online Courses and co-hosts The Data Journalism Podcast. How Charts Lie
Cairo’s light writing style and focus on the general reader make this a quick read. The many examples increase the immediacy and relevance of the material. Numerous charts reinforce his points, and his writing is packed with valuable insights.
The “Introduction” motivates the reader with examples of 2016 US election maps, US murder charts, and expiring tax cuts to show how charts and maps can mislead. These highlight the importance of understanding what is counted, the impact of timeframes, and the effect of design choices on the message of a chart.
Any chart, no matter how well designed, will mislead us if we don’t pay attention.
Once the reader understands how charts can lie, Cairo reveals the elements of a chart and how to read them. “How Charts Work” lays the foundation for the book. Cairo identifies two features of a chart: the scaffolding and the content. The scaffolding includes everything but the data itself; the title, subtitle, units of measurement, axis labels, and data source, along with annotation or short textual notes.
The chart content refers to the visual encoding of the data, i.e., the length, height, position, angle, color hue, or color shade of the symbols associated with the data. Not only does Cairo describe the visual encodings, but he also discusses the appropriate use of symbols based on the underlying data. He reviews their use in different types of charts with tips and tricks for quickly understanding the charts.
Cairo discusses the basic chart forms, such as bar graphs, line graphs, scatterplots, and maps, as well as more advanced forms like treemaps, heatmaps, parallel coordinates plots, and connected scatterplots. He concludes the chapter with a 5-step process for reading charts. A quick glance at a chart can provide insight that can be followed by a more detailed reading.
Given a motivated reader who understands the elements of a chart, Cairo identifies five specific ways that charts can mislead.
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 “Charts That Lie by Being Poorly Designed,” Cairo discusses potential issues with dual-axis charts that purport to show the relationship of two variables (using Planned Parenthood data), how 3D effects can distort the perception of data (using fictional company sales), problems comparing two variables that are charted with different scales (using unemployment data), and the importance of axis scaling on the perception of chart slope (using world life expectancy data).
Each of the “Charts that Lie …” chapters contain multiple well-illustrated examples, often supplemented with summary quotes. The reader will see classic, new, and fictional examples of misleading and/or misinterpreted charts.
Cairo not only shows examples of misleading charts, but offers alternatives that are less misleading, given the purpose of the chart. In many cases, the better choice is obvious. In some cases, there is not a single, clear-cut solution, but there may be better or worse choices.
While there are too many examples to describe each one, Cairo’s description of the hurricane cone of certainty is particularly noteworthy. Experts on weather, climate, and environmental science note that “nearly everyone who sees this map reads it wrong.” He not only explains why the chart is misread, but he also describes the purpose of the chart and how to read it properly. He then identifies additional charts that could be used to help the reader better understand hurricane forecasts and their uncertainty.
Cairo uses summary quotes to highlight important ideas for the reader. One of his more famous quotes is:
… we must stick to the principle that a chart shows only what it shows and therefore, we must strive not to read too much into it.
This is especially relevant for scatterplots that show a strong correlation. The reader must be alert not to assume that correlation implies causation…additional evidence is needed!
The book’s “Conclusion: Don’t Lie to Yourself (or to Others) with Charts” uses one of Cairo’s favorite charts, Florence Nightingale’s “wedges” chart of causes of mortality in the Crimean War. The chart shows the large proportion of unnecessary deaths during the Crimean War as a result of preventable infections, as compared with combat deaths, and the dramatic reduction in deaths with the introduction of sanitary measures.
Cairo identifies three important principles that can be learned from the chart:
For a chart to be trustworthy, it must be based on reliable data.
A chart can be a visual argument, but it’s rarely sufficient on its own.
Data and charts can save lives and change minds.
The “Afterword,” which is new to the Second Edition, brings the book’s lessons together, with an analysis of COVID-19 pandemic charts. He uses them to highlight his earlier content, discussing subjects such as the importance of knowing what is counted, i.e., the definition of a COVID-19 death and the definition of excess deaths, understanding the uncertainty of the data, and the use of logarithmic scales to reveal rates of change.
How Charts Lie: Getting Smarter about Visual Information is an excellent, all-purpose introduction to chart reading and the ways that charts can mislead. The chapter “How Charts Work” is a succinct yet comprehensive primer on chart reading, and his “Charts That Lie …” chapters are packed with examples to help the reader become a more critical chart reader.
Rather than always blaming the chart maker, Cairo understands that even well-intentioned charts may mislead. Cairo chose his title How Charts Lie carefully, focusing on the object, charts, rather than the maker. This contrasts How to Lie with Statistics and How to Lie with Maps, which focus on the act of lying.
This is the second review in our “lying” series of book reviews. Next up is How to Lie with Maps, by Mark Monmonier.
Resources
How Charts Lie: Getting Smarter about Visual Information publisher’s website.
How Charts Lie: Getting Smarter about Visual Information Amazon website.
How Charts Lie: Getting Smarter about Visual Information YouTube presentation.
Alberto Cairo profile from the University of Miami School of Communications
Alberto Cairo’s Weblog about design and visualization
The Data Journalism Podcast with Alberto Cairo and Simon Rogers
댓글