Two phenomena have come to rule social media in recent years like few others have. The first is, of course, disinformation, which tries to spin and sway your mind towards doctored narratives. The second, an unlikely one, is charts with data. Graphs that many detested in maths classes have suddenly turned ubiquitous, making numbers fashionable through cool maps (and even boring bar charts with make-up applied).
Unfortunately, the two phenomena intersect way too often, and feed on each other’s popularity. That’s where Kavitha Ranganathan’s book, Impactful Data Visualization: Hide and Seek with Graphs, comes to the rescue.
It doesn’t take a nerd to understand a chart: that’s why the good ones go viral every other day. But neither does it take a novice to misunderstand a chart. Bad ones, too, go viral, without the cleverest of minds noticing the defects. Charts can hoodwink us in numerous ways, and Ranganathan’s book explains the nuances in simple language with lots of visual, easy-to-follow examples.
Can you, for example, spot the sin in a chart showing you two bars, one visibly double the other in length—except that the numbers they represent are, in fact, as close as 110 and 120? Political parties and sympathetic media outlets have often used this strategy across the world to misguide voters. Or, what about a line chart that shows a sudden jump in a company’s revenues—except that the left half of the timeline tracks revenues at quarterly gaps and the right half at annual gaps?
An explanation to all this makes the book an important crisp primer for both professional chart-makers (read, nearly everyone in this era of corporate meetings and slideshows) and lay readers (again, read, nearly everyone with an internet connection) alike. In that, the book seems inaptly named: a book called Impactful Data Visualization sounds more handy for chart-makers only, but this one’s more. Also, the book isn’t as much about “impactful” charts as it is about truthful, honest and accurate charts.
Ranganathan’s book was published by Sage Publications last year, and has now been published by Penguin Random House India as part of its IIM Ahmedabad Business Books series. Ranganathan, a faculty member at the business school, has taught data visualization for years, catering to various kinds of audiences.
The book starts by laying down a list of all reasons we ever use charts for—comparisons (such as bar or column charts), time trends (like line charts), correlations (such as scatter plots), and composition (for example, pie charts or stacked bars) among others. This useful list brings a system to the science of making charts, given that many beginners get baffled when asked to choose among dozens of chart options that Microsoft Excel provides for any given dataset.
The book then gets into the common traits of untruthful charts (such as wicked choices of scale—the first example mentioned above is possible when the Y-axis starts at 100 instead of zero, giving the illusion that the metric has doubled in size between 110 and 120), particularly important in this age of disinformation. There’s a chapter on common pitfalls of well-meaning charts that may fail to make a point or may be deceptive in deeper ways. One chapter explains design choices, or the art of chart-making: colour palettes, position of text, when to use bold text and when not to, and so on. The last chapter brings it all together to explain how to ‘tell your data story’.
Charts are made to make communication easier, not to muddle it up. To that end, you know by the end of the book why both the art and science matter in ensuring a chart does its job. This is important because many beginners in data tend to see pedantic colour and design choices as their bosses’ whims, when they are, in fact, conscious choices meant to aid communication.
The book cites ample research to show proven evidence from human experiments of why and how a certain visual trick helps when you are reading a chart. Research adds to the juice that the book brings—for example, did you know that even in a visual, it’s the textual part (such as the chart headline) that you are the most likely to recall later? Or that there are specific cases where the otherwise notorious pie charts are actually helpful, and that this was shown through a 1991 study?
But there’s a flip side: the book, in parts, relies more than needed on quotes and views expressed by prominent statisticians in their own books—at times making the text sound like a literature review when Ranganathan could have made the book fully her own. When she doesn’t lean on the work of others, she’s convincing and insightful enough. This flaw is more about the textbook-like tone and flow of the book, when some other books about data and numbers published in recent years have adopted a more narrative-driven approach.
There’s another area the book could have done better. It has multiple examples of its own, shown visually (many in the before-and-after format), but where it cites real-world consequences of flawed chart-making, it mysteriously uses only text, breaking the smooth flow. Web addresses of bad charts by prominent organizations and media outlets are given as footnotes: but those examples are essential to understanding the flow of what Ranganathan is trying to explain. The explanation of a misleading Reuters chart on gun deaths in Florida even throws the reader back into an older “well-known” graph on US military deaths in Iraq—neither chart is shown and the reader is expected to understand a whole page of visually descriptive text through the cited URLs. Visual renditions of the flawed charts could have helped explain the problems better—and quicker, without one having to open the link on the side.
The easy availability of chart-making tools means the activity is no longer too niche to ignore. If your day job needs you to make PowerPoint slides to captivate funders, customers, bosses, or students, you most likely make charts in some way. At Mint, we publish nearly two dozen every day in an effort to simplify the crux of complex stories. On the flip side, if you are a funder, a customer, a boss, a student, or a Mint reader, you most likely see charts in some way, often for a fleeting moment in which you need to make a conclusion or decision. You may also have the urge to forward charts about your favourite political party or football team on WhatsApp. The wolves are out there to deceive you, but so are many innocent minds who may not realize how their attempt at a chart could be unintentionally deceiving. By the end of Ranganathan’s book, you will surely know the grammar of making honest charts and also be vaccinated against the dishonest ones.