Once Upon A Datum

Learning the magic of data-driven storytelling.

Photo by Brian McGowan from Unsplash.

Who hasn’t struggled with telling an effective story? Storytelling is challenging to get right, and being a great storyteller is a trait possessed by few and desired by many. There are several challenges: where to begin, what parts to include (and what to exclude), and how to keep a coherent storyline. Well, guess what’s even harder? When you have to stick to actual facts. In other words, data-driven storytelling.

This is the challenge outlined by Yang et al. in a recent paper that adopts the famous Freytag’s Pyramid structure, a traditional paradigm of dramatic structure, to data-driven storytelling using visualization. In the paper, the authors begin by condensing the pyramid into the following steps:

1. Setting: Gives introductory information about the data story and grabs the audience’s attention.
2. Rising-Climax: Builds the tension of the story with supporting facts towards the climax by presenting core insights.
3. Resolution: Provides conclusions and take-away messages.

These steps are an excellent start to structuring your story. However, what is the best way to create a data-focused narrative throughout your entire set of visualizations? To answer this, the authors present a framework based on three dimensions of narrative visualizations for telling stories with data:

• Narrative pattern: Provides narrative devices as initial ideas to assist the creative process.
• Data flow: Strategies for selecting, organizing, and formalizing individual story pieces (i.e., facts backed up by data).
• Visual communication: Supports strategically selecting visual representations.

Using these dimensions, Yang et al. present a design space for the above three steps:

Design Space Graph by Yang et al.

Setting

Introducing visualizations: Guide the audience in how to read the graphs:

• Data Flow: What is an effective way to present information?
• Visual Communication: How should you explain the context of the data behind each visualization?

Statistic Hook: Foreshadowing the data story:

• Data Flow: What significant data stands out and will grab the viewer’s attention?
• Visual Communication: How can you build up to reveal the statistically significant information?

Preview: Gives a hint in advance of what to anticipate

• Data Flow
  - Can change over time be shown with time-series data?
• Visual Communication
  - Can the visualization be fast-forwarded or quickly compared?

Raising a question: Directly ask the viewer a question:

• Data Flow: What questions about the climax can be presented?
• Visual Communication: How should you emphasize this information?

Introducing backgrounds: Contextual information about the story:

• Data Flow: How is the data collected? From whom? Why collect it?
• Visual Communication: What illustrations help to convey abstract information?

Presenting concrete characters: Make the story personable for viewers

• Data Flow: What data facts directly affect the characters?
• Visual Communication: What do the characters look like?

Rising-Climax

Showing contrast: Create a plot turn to setup the climax:

• Data Flow: What are data comparisons that can be made?
• Visual Communication: How can juxtaposition and visual tricks build tension?

Showing accumulative significance: Intensify the story with similar data facts:

• Data Flow: How can you present similar data from different perspectives?
• Visual Communication: How can facts be staged to build tension?

Showing the decisive moment: Bring the viewers to a important moment

• Data Flow: What shows change over time?
• Visual Communication: How can feelings of excitement or anticipation be enhanced?

Showing ranking: Establish suspense by revealing things one-by-one:

• Data Flow: How can data facts be ranked by importance?
• Visual Communication: How can you group data effectively in a row or set?

Resolution

Recap: Remind audience of main messages within the story:

• Data Flow: What are the indispensable data facts?
• Visual Communication: How can findings be summarized?

Predicting the future: Give a glimpse into the future:

• Data Flow: What is the transition of data from present to future?
• Visual Communication: How can new data be emphasized and compared to historical data?

Echoing the beginning: Revisit story context from the beginning:

• Data Flow: How can data shown at the beginning be effective at the end?
• Visual Communication: What elements from the beginning can be reused at the end?

Next steps: Solutions and actions viewers can take:

• Data Flow: What external domain knowledge can be included?
• Visual Communication: How can you show solutions and aspirational data?

Summary

Telling a story with data is not a magical or mystical process that involves throwing graphs at an infographic to see what sticks. Rather, telling data visualization stories is about understanding how your data should be structured to best tell your reader what you want them to know. Instead of feeling stressed about this unique form of storytelling, utilize this design space and guideline to incrementally improve your storytelling skills. The only thing that’s magic about this process is the realization that readers can truly become engaged in a compelling and well structured story about numbers.

This is a summation of the academic paper, “A Design Space for Applying the Freytag’s Pyramid Structure to Data Stories”; see full citation below.

References

Yang, L., Xu, X., Lan, X., Liu, Z., Guo, S., Shi, Y., & Cao, N. (2021). A Design Space for Applying the Freytag’s Pyramid Structure to Data Stories. IEEE Transactions on Visualization and Computer Graphics, 28(1), 922–932. https://ieeexplore.ieee.org/abstract/document/9552203