Visualization & Story

From FASL and Syfr Learning LLC

Remember data-driven decision making? It was a great idea but its future was doomed from the beginning. Let’s look at a quick example to explain why. Read the paragraph and answer the question before going further. 

Maria is 30 years old, single, outspoken, has a degree in with majors in philosophy and psychology. She graduated cum laude from a prestigious liberal arts university and was deeply involved in social justice issues in college. Which of the following alternatives is more probable?

Maria works as an investment banker.

Maria works as an investment banker and is very active in the Progressive Movement within the Democratic Party.

This is slightly updated from the experiment Daniel Kahneman used about 30 years ago but the structure and the choices are essentially the same. When Kahneman gave this question to undergraduates, 85% to 90% chose option two. Why? When the introductory paragraph was being read, the brain created a story about Maria that made sense – it was coherent. Becoming an investment banker did not exactly fit the story but being active in the Progressive Movement did. So, 85% or more of these unsuspecting undergraduates followed their story. 

But, option two is incorrect. Option two is a subset of option one. You know that Maria is an investment banker because that is part of both options. To pick option two is to take a risk on a much, much smaller subset. So, let’s go back to why.

There may be many reasons for preferring option two but we suggest three. 

  1. Statistical reasoning is not intuitive. We can name only one superintendent who was a math major. So, not only are humans in general not statistical thinkers, few educational leaders majored in math or even liked statistics. 
  2. Stories are intuitive. Humans create stories to explain why. When the stories are coherent, they are powerful. They trump data.
  3. Our brains easily generalize from the specific. It is far more difficult to go from the general to the specific. 

How can we compensate? One way is to make the data visual and visualization has become almost essential when it comes to comprehending big data. Our favorite visualizer of data is Hans Rosling.

Rosling was a public health professor in Sweden who had worked in Africa for extended periods of time. For as long as he could remember countries had been described with labels like “developed” and “undeveloped” or “developing”. A country existed in one of the two. Today, however, 85% of the world’s population lives in developed countries by the definitions of 1965. Most of the rest are rapidly approaching ‘developed’, yet few people actually believed it. 

With a problem like this, Rosling needed a better solution than raw data and he decided to make the data visible. You can see one way that he did it here: Even with this kind of persuasive, visual presentation, he sometimes lost his audience. His next solution was a website that illustrated the progress with pictures. For example, one of the first signs of growing income is footwear. It goes from bare feet to flip flops to various versions of sneakers. Each representing a change in income ( The website is far more likely to create a story in the viewer’s head but still, the job of changing someone’s mind was daunting.

There is a classic story of data visualization told in Steven Johnson’s book, Ghost Map. In 1856 cholera struck London and within a few weeks had killed thousands. A doctor, John Snow, developed a theory that cholera was a waterborne disease and created an elaborate map that documented every cholera case in a specific area of London and tied all of them to one water pump. It was a pump that drew water from a well infested with human waste. Because cholera tended to occur in areas that smelled bad, because of waste, it was assumed that it was an airborne disease and Snow’s map was dismissed.

About a decade after Snow’s death, there was another outbreak of cholera. Again, it was localized around water pumps, Snow’s map was brought out and he won a partial battle posthumously. His theory about the cause of cholera was still disputed but London, and quickly the western world, began to install sewers and soon thereafter water treatment plants. The modern city was essentially born then – it could grow. While visual data did it, it was a tough sell and Snow did not live to see it. 

Visualization’s importance extends beyond data. John Hattie’s work has a brilliant title: Visible Learning. Make learning visible, most importantly for students and teachers. Use graphs, color, pictures, anything that drives a visual representation of learning. It will become the core for teachers and students to internalize a sense of learning progress.