Some people say a data-driven decision is objective and based on facts. However, the way of processing data or how visualizing the data might be subjective, which involves an author’s intention, or maybe just a mistake. In this posting, I’d like to introduce common mistakes or calculated distortion, found in real-life visualizations.
Non-zero baselines
First of all, an error of visualization is raised when the y-axis of the bar chart does not start from 0 (zero). In the example of Fox News, showing the changes in tax rate if the tax cuts of the Bush administration expire, the result seems to be dramatic. However, seeing the y-axis which starts from non-zero, the change presented by Fox News seems significant.
If the y-axis of the bar chart starts from 0, the change seems trivial. Here’s the comparison of two bar charts with different y-axis.
Considering the fact that Fox News is a pro-republican, the distortion of the visualization might be intended to exaggerate the effect of Bush’s policy.
Scale Misrepresentation
Another example of misleading visualization is the manipulation of scale. The below plot from The Economist shows the GDP per person of the poorest and richest regions. The maximum value of the X-axis is 400, but Britain’s values, 848 and 1,150, are beyond the maximum. They are more than two times bigger than 400, but the length of the charts is truncated, barely exceeding the maximum label, 400.
The correct way of visualizing lengths, which are proportional to values, would be as follows. The re-drawn visualization effectively shows gaps between the poor and the rich in Britain. The gap is much bigger than those of the other countries.
Cumulative Plot
In Apple Inc.’s conference call, Tim Cook presented cumulative iPhone sales, indicating a dramatic increase since its release. Based on the visualization showing an upward trend of cumulative sales, investors might have had a positive impression of the result at first.
However, the bar charts of quarterly sales, which are re-drawn by the numbers from earning reports, explicitly show the downward trend of recent quarters.
The visualization from Apple is based on the true number, but Tim Cook would have been intended to disguise declining sales.
Colors
Colors may imply different meanings in different countries. In visualization, authors should consider the cultural background of viewers since the same color could suggest totally different meanings. For example, in most countries, red indicates a decrease in stock price whereas green represents an increase. However, in the Korean stock market, red stands for an increase in price while an increase is symbolized in blue.
In the candlesticks of Samsung Electronics from two different websites, one from Naver and the other from Yahoo Finance, the colors of the daily candlestick are reversed. Though not intended, the colors used in the visualizations could mislead viewers from different countries.
Reference
Cairo, A. (2020). How Charts Lie: Getting Smarter about Visual Information (1st ed.). W. W. Norton & Company.