How Charts Lie?
🌀 Key Takeaways from Alberto Cairo's Book
One of my favorite data visualization books is Alberto Cairo's How Charts Lie. It offers great insights into the charts you create and highlights the importance of making them clear and effective. Here are some key takeaways I highlighted for myself, but if you have the chance to read it, definitely do so! It's very interesting and easy to read, and most importantly, it helps develop visual literacy and critical thinking about charts.
“A picture is worth a thousand words.” My hope is that you’ll soon stop using it unless you append it with this: “if you know how to read it.”
Alberto Cairo. “How Charts Lie”.
Charts can be misleading for several reasons: they may show incorrect or insufficient data, be poorly designed, or present information that isn't truly there. The quality of the data plays a significant role in the effectiveness of a graph. Any graph can deceive if viewed without careful consideration. Additionally, charts can be manipulated to support any argument. However, well-designed charts help us make informed and effective decisions.
We often use charts to reinforce our beliefs and become more convinced of them. If we have a preconceived idea and then encounter a graph that appears to support it, we're likely to accept it blindly and become even more convinced. For example, consider the charts related to COVID-19: many have been used to either confirm or refute the effectiveness of masks, the number of deaths, and other related issues.
“Charts can be instruments for either reasoning or rationalization.”
You need to learn how to interpret graphs effectively and become more graphicate to read information from graphs accurately. The text also points out that just looking at an essay isn't enough to understand its content fully. Similarly, many graphs require careful examination and analysis to avoid being misled.
“We all need to turn into attentive and informed chart readers. We must become more graphicate.”
You need to critically examine the data presented in charts, as the data can often be misleading. First, check the source to determine if it's reliable and whether you can verify the data or replicate similar results. If the source is not provided, it's best to question the credibility of the chart.
“Being an attentive chart reader means being a critical reader of data.”
A chart only shows what it shows. They are all often simplifications of reality, so it can show a lot, but it can also hide a lot.
“A chart shows only what it shows, and nothing else.”
Don't try to extract conclusions from a graph or data that aren't supported by the information presented. The same data, when viewed from different angles (e.g., with different filters, manipulations, or contextualizations), can produce vastly different results. This is especially true for graphs that align with your existing beliefs—such cases require even more critical scrutiny. Always remember that correlation does not imply causation. For examples of misleading correlations, you can visit the site Suspicious Correlations. As the saying goes, “If you torture data long enough, it will confess to anything.”
Many data points are meaningful only in the context of other data. For example, comparing wages or costs between countries without considering purchasing power can be misleading. Percentages can also oversimplify, as we may not know the specific numbers behind them. For instance, 1% of the world’s population may seem small, but that represents over 60 million people—about 12 times the population of Finland!
“Different levels of thinking may require different levels of data aggregation.”
The book contains many interesting examples. Many of these them involve political stories about voting or cherry-picked data related to ecology and population, so it's important to be extra cautious. We don’t always have time to delve deeply into every graph, which is why, in graphs intended for a broad audience, it's crucial to carefully examine what message people might take away and its accuracy. Additionally, it's important to include disclaimers and sources directly with the graphs to clearly convey what the graph represents.