Lately, I've stumbled across a number of blogs discussing how data visualization can be used to tell stories.
"If it's one thing every good story has, it's conflict."
Depending on who you speak to, the concept of storytelling can be either very definitive or ambiguous. Out of all the interpretations I came across, the one typically used by most screenwriters, playwrights and authors stuck with me the most. For the sake of this blog, we'll use their explanation of what a story is and see how data visualizations can adopt this model to tell stories.
Storytelling: the screenwriter's perception
If it's one thing every good story has, it's conflict. No one wants to hear a story about a person who achieves his or her goal with little to no resistance from another character (or characters). There's a reason why "The Shawshank Redemption" is such an appealing story – the protagonist has a lofty goal (that is, to escape from prison) and everything's working against him (the law, the warden, the guards, etc.)
Now that we've got that out of the way, here are the other essentials to a good story, with examples:
- Character vs. characterization: The protagonist isn't actually what he or she appears to be. On the outside, Bruce Wayne appears to be a billionaire party animal, but he's actually a vigilante with an unbreakable moral code.
- Want: The protagonist must have a desire that he or she spends the entire story trying to fulfill. A high school student does everything he or she can to get into college.
- Theme: An underlying message that is tied to the protagonist's journey. The theme behind "Beauty and the Beast" is that not everyone is the way they appear to be.
In addition, a story must inspire empathy, or a feeling of understanding, in the audience for the protagonist.
Telling stories with data visualization
Obviously, a data visualization that intends to tell a story isn't going to fit the same criteria as an award-winning film. However, there are ways in which data scientists and other professionals working with visual analytics tools can use that model as a set of guidelines.
Before providing you with an example of a visualization that tells a story, it's important to translate the essentials highlighted above into factors that can be assembled to create a concrete data analysis:
- Protagonist is the subject: The subject can be a market segment, industry trend, factory, supply chain or any other kind of business concern you could think of. Through the data you collected about your "character" you can set rules in a data visualization program that assign your subject behaviors – i.e. a factory would react poorly to a broken machine.
- Associated data is the conflict: Outside elements will influence the way in which a subject reaches its goal. If a car factory "wants" to produce 500 sedans per day, internal and external factors may either hinder or support this intention.
- The plot as a time sequence: If you want to mimic the feel of a film, use a time series that details the progression of your subject's journey, watching how your subject's path is either riddled with conflict or smooth as silk.
"How do you inspire emotion in the people who are viewing your visualization? "
But how do you create empathy? How do you inspire emotion in the people who are viewing your visualization? This is where color, shape and movement come into play. A study by Pennsylvania State University found that horizontal lines inspire a sense of tranquility, while vertical lines relate to strength, so pick and choose depending on the type of story you're trying to convey.
As for colors, Color Wheel Pro defined several meanings that can help you manipulate the emotions of your audience. For example, orange is associated with joy, fascination and enthusiasm.
So, are storytelling and data visualization a match made in Heaven? Again, it depends on who you speak to, but the examples highlighted above would suggest that they're compatible.