Fundamentals of Data Visualization

🌈 Abstract

The article discusses strategies for effectively communicating insights from data visualizations by framing them as compelling stories. It covers the importance of storytelling in data visualization, the key elements of a story, and techniques for creating memorable and clear figures that support the narrative.

🙋 Q&A

[01] Telling a story and making a point

1. What is the importance of storytelling in data visualization?

• Data visualization is often done for communication, so presenting insights in the form of a clear and engaging story is crucial.
• Storytelling helps get the audience interested and excited about the data, rather than leaving them bored or confused.
• If a clear story is not provided, the audience will make up their own story, which may be inaccurate or undesirable.

2. What are the key elements of a story?

• A story is a set of observations, facts, or events presented in a specific order to create an emotional reaction in the audience.
• Good stories have a clear story arc, with a build-up of tension at the beginning followed by a resolution.
• Common story formats include Opening-Challenge-Action-Resolution, Lead-Development-Resolution, and Action-Background-Development-Climax-Ending.

3. How can data visualizations be incorporated into a story?

• A single static visualization rarely conveys an entire story. Multiple visualizations are usually needed to cover the different parts of the story arc (e.g. opening, challenge, action, resolution).
• It is also possible to condense an entire story arc into a single figure, but this requires including both a challenge and a resolution.

[02] Make a figure for the generals

1. What is the "make a figure for the generals" concept?

• The audience, even if they are experts, may not be able to rapidly process complex visualizations. Simplify figures as much as possible, removing anything tangential to the key story.
• Generals (e.g. busy decision-makers) should be able to immediately understand the main point of each figure.
• Scientists are often not trained to make figures for this type of audience, instead creating overly complex visualizations.

2. How does the ease of modern visualization software contribute to this problem?

• The ability to create sophisticated, multi-faceted visualizations can tempt people to include too much information, making the figures hard to understand.
• A simpler visualization like a bar graph may be more effective at conveying the key insight than a complex figure.

[03] Build up towards complex figures

1. What is the recommendation for introducing complex figures?

• When using complex figures, first show a simplified version to help the audience understand the key elements before presenting the full, detailed visualization.
• This is particularly relevant for small multiples plots, where showing a single subplot first makes the full grid easier to digest.

[04] Make your figures memorable

1. What is the balance to strike between simplicity and memorability in figures?

• Simple, clear figures avoid distractions but can end up looking generic and forgettable.
• More visually complex and unique figures are more memorable, but this complexity should not hinder the audience's ability to quickly understand the information.
• Isotype plots, where repeated images are used to represent data values, can make figures more memorable while still conveying the information clearly.

[05] Be consistent but don't be repetitive

1. What is the recommendation for using consistent visual language across multiple figures?

• Figures describing different analyses should look visually distinct, using different visualization approaches, so the audience can easily recognize where one analysis ends and another begins.
• However, the figures should still have a consistent overall visual language so they appear to belong together.
• Avoid using the same type of visualization for multiple parts of a story, as this can become repetitive and obscure the main points.

2. What is the recommendation for the progression of figures in a story?

• Start with figures showing the raw data, then move to more derived quantities (e.g. percent changes, averages) in subsequent figures.
• The number of figures should generally be limited to 3-6 per distinct story or subplot, to avoid overwhelming the audience.