Storytelling with Data by Cole Nussbaumer Knaflic is an essential resource for data visualisation professionals, clients and everyone looking to derive insights from data.
This post provided a detailed summary and insights from each chapter and my overall thoughts.
As a Data Analyst working with BI tools such as Excel, Power BI and Sisense, I have developed several reports for clients across multiple sectors (one made it to the TV during the COVID-19 pandemic). Whilst I got lots of positive reviews from clients and colleagues, I realised captivating my audience was an issue. Multiple factors such as choosing the right visual, colour, alignment and filling up space continued to linger in my mind. I was good at developing the report functions but struggled with the design. That lead to my interest in reading Storytelling with Data.
Chapter 1: The importance of context
Exploratory vs Explanatory Analysis
Exploratory Analysis guides the user to understand the data and interesting points to highlight to an audience.
In contrast, the explanatory analysis turns the data into insights to be consumed by the audience, which avoids showing everything in the data. The focus is on Who, What and How
WHO: Target audience, e.g. doctors, students, board of directors.
WHAT: Knowledge your audience should know. Tell the data story to the audience with a focus on critical interests.
HOW: Mechanism of communication. Avoid cognitive load; consider your level of control in different scenarios. In a live presentation, you are in control. In a written document or email, your audience is.
The 3 min story: Present points concisely within an optimised time slot. Express the big idea with short sentences and action words. Focus on problem identification, problem-solving, problem feedback.
Storyboarding: This is a significant part of addressing points during communication. It is a visual roadmap of the steps required to structure communication flow. Starting with a sticky or cards will clarify how the data should be told.
Chapter 2: Choosing an effective visual
Most effective visuals
Simple Text
Two-point comparisons aren't effective in showing variation. Instead, consider a card highlighting the difference in a simple text format. E.g. Instead of showing a percentage increase in people vaccinated in 2021 and 2022 (overall), show the text as a dynamic value as “ There is a 30% increase in vaccinated population in 2021 compared to 2022.
Tables
Effective for communicating with a mixed audience. It's advisable to avoid tables for live presentations.
A use case would be for visualising multiple units of measure across different departments.
Key: Avoid heavy borders and lines on tables; light or no boundaries are preferable.
Align values to the centre stage when presenting a table.
Heat map (Conditional Formatted Table):
Leveraging colours to highlight the magnitude of numbers. E.g. Setting a KPI and indicting increase or decrease in targets on a colour gradient.
Graph
Preferred method of visualising data. Frequently used:
Point - Scatter Plot: Observe the correlation (relationship) between two things. Use case: identifying the correlation between COVID-19 lockdowns and HIV service disruptions. An effective way to further utilise the scatter plot is through colour, i.e. colours above average are shaded differently to buttress variation.
Line: Preferable for time series analysis. An area line graph can contain two or three series along the x-axis, e.g. max sales, average sales and minimum sales.
Bar Chart: Traditionally more popular, therefore easier to understand. Bar charts need a zero baseline on both the x and y-axis. Bars can have single or multiple series. Horizontal bar charts are preferred for categories with long names since data is processed from left to right.
Slope graph: Easy to express much information about the categorical difference between two or three time periods. Most visualisation tools can be tricky to access; hence, patience is required. The lines connect show rates of difference in an intuitive manner.
Visuals to avoid:
Pie Chart
Donut chart
Secondary y-axis and 3D visuals
Chapter 3: Clutter is your enemy
Cognitive load
Understanding data visualisation at a glance can be overwhelmingly busy and complicated if the critical information is not highlighted correctly.
Imagine a computer working with dependencies on the computer processing power to function. A similar process occurs an audience is asked to process data visually. Therefore, it is essential to reduce unessential information for faster processing by the target audience.
Clutter
These are visual elements that take us space but do not increase understanding. They lead to a bad load cognitive load.
Key: Reduce Visuals.
Knowing the visual objectives deliverables helps reduce clutters.
Gestalt Principles of Visual Perception
Aids in communicating the information needed and avoids clutter.
There are six (6) principles:
1. Proximity: Think of objects physically close together. To see groupings as columns and rows.
2. Similarity: Objects similar in colours, shapes, size or orientations are perceived as related to the part of a group.
3. Enclosure: Think of objects that are physically enclosed together as belonging to part of a group. Light shading is often enough. Highlight difference with shading—E.g. Shading Forecast from actual in a time-series analysis.
4. Closure: People like things to be fit and straightforward. Removing unnecessary elements in a graph makes the data stand out. Borders and background shadings are not necessary for a visual.
5. Continuity: This principle is similar to closure; our eyes try to seek the smoothest path and create continuity in what we see even when it may not explicitly exist. Your visuals should show continuity for easy understanding.
6. Connection: Helps to create a from different points in a visual.
Key:
Avoid Lack of visual order, group related items in the same colour
Align text to either left or right justify
White spaces are essential in helping the audience understand a crucial point.
Avoid non-strategic use of contrasting colours and shapes.
Use case: A line chart that shows variation in the application received vs processed.
Remove chart border
Remove gridlines
Remove data markers
Clean up axis labels(E.g. January to Jan, 1,000,000 to 100 with the legend specifying…in millions)
Label data directly
Use a consistent colour
Chapter 4: Focus your audience's attention
Leveraging preattentive attributes to direct audience attention and visual hierarchy to control information processing.
Roles of the brain in visual communication design:
Iconic memory: Happens super fast with our consciously realising it. Information stays in your iconic memory for a fraction of seconds before it moves to short term memory—tunes information into preattentive attributes.
Short-term memory: Typically with limitations because about four (4) chunks of visual information can be stored at a given time. It is easy to lose the audience with a cognitive load where a visual has multiple colours, shapes and data makers. This can be solved by labelling the data series to minimise back and forth between legend and data. It also fits larger, coherent chunks of information into a finite space in the audience’s working memory. In conclusion, when information leaves short term memory, it is lost forever or got to the long-term memory.
Long-term memory: is built up over a lifetime and is vitally important for pattern recognition and general cognitive processing. Visual and verbal memory summarises that cat differently.
Pre-attentive attributes: where to look?
Text: highlight focus area based on colours, size, outline, bold, italics, underline.
Graphs: highlight parts of the bar and text to hone the audience focus.
Size: Relative size denotes relative importance. Size important highlights bigger.
Colour: As the most powerful tool for drawing the audience’s attention, use colour consistently design with colourblind in mind. Be thoughtful of the tone colour conveys. When selecting brand colours, two colours sound be sufficient, preferably one.
Position on Page: Specifically, audiences look at information from top left to top right, then bottom left to bottom right. Focus the audience’s attention on where you want it on the page.
Chapter 5: Think like a designer
Form follows functions: What do you want the audience to do with the data (process)? Create the visualisation (paper) for this.
To achieve perfection in data visualisation, avoiding is more critical than including.
Not all data are equally important
When detail isn’t needed, summarise
Ask yourself: would eliminating that change anything?
Push necessary non-message impacting items to the background
Create a clear visual hierarchy of information
Accessibility
Don't overcomplicate - if it is difficult to read, it is difficult to use.
Text is your friend
Aesthetics
Acceptance - find ways to get your audience to adapt your design.
Chapter 6: Dissecting model visuals
How do you emphasise and deemphasise audience attention? Focus on building intentional visuals regardless of the model applied.
Model visuals
Line Graph: Round axis labels. There is no correct way to represent day, weeks, year; instead, focus on what and how you want your audience to use the visual and make a deliberate decision in light of things.
Annotated line graph with fading and distinction between actual and forecast
100% stacked Bar charts: focus audience attention using colour
Leverage positive and negative stacked bar charts
Horizontal stacked bars: show and group (using colours) top priority at the top
Chapter 7: Lessons in storytelling
What exactly is a story?
At a fundamental level, a story expresses how and why life changes. Stories start with balance. Then something happens—an event that throws things out of balance. Key: Keep the information simple. Edit ruthlessly. Be authentic. Don’t communicate for yourself—communicate for your audience. The story is not for you; the story is for them.
Think about the beginning, middle and end. Make headlines first, so the audience knows what to follow.
Narrative flow - the order of the story
Think of each audience member and how best you can get them to engage with the visualisation. The data story must have a flowing, compelling and robust narrative. The narrative flow is the spoken and written path along which you take your audience over the course of your presentation or communication. This path should be clear to you. If it isn’t, there certainly isn’t a way to make it clear to your audience. One way to order the story—the one that typically comes most naturally—is chronological.
Leverage repetition
leverage the power of repetition by telling the audience what you will cover, covering much content, and summarising the report. That way, the information is cemented in the audience’s memory.
Implement tactics to help ensure that your story is clear
Horizontal logic: Have an executive summary slide up front, with each bullet corresponding to a subsequent slide title in the same order. This follows the repetition first step.
Vertical logic: ensures all information on a given slide is self-reinforcing.
Reverse storyboarding
It takes the final communication, flips through it, and writes down each report page’s central point. If reverse storyboarding does not help you understand structurally where you might want to add, remove, or move pieces around to create the overall flow and structure for the story that you’re interested in conveying
A fresh perspective Share reports and visuals with a friend or colleague. It can be someone without any context, ask them for their feedback and what they paid attention to. It is by making them understand the data to show relevant insights.
Final Chapters: Pulling it all together and case studies
The last three chapters (chapter 8: pulling it all together, chapter 9: case studies and chapter 10: final thoughts) focus on tying together the previous seven (7) chapters. The goal followed step by step was to craft a compelling narrative and storytelling.
The case studies discuss considerations and solutions for tackling several common challenges when communicating visually with data. The big lesson here is that you have several alternatives to pies that can be more effective for getting your point across. Key: When you find yourself in a situation where you are unsure how to proceed, I nearly always recommend the same strategy: pause to consider your audience. The responsibility —and the opportunity—to tell a story with data is yours. Develop a stylistic approach that makes data easy to understand.
Further case studies can be found on storytellingwithdata.com
Final tips for maximising visualisation success:
Learn your tools well
Iterate and seek feedback
Devote time to storytelling with data
Seek inspiration through good examples
Have fun and find your style
My Summary and Thoughts
I absolutely recommend Storytelling telling with Data as base inspiration for data visualisation. If you want to make clients and colleagues understand the value of well-designed graphs, charts, and reports, it will be a fantastic resource. It is excellent at explaining relatively complex concepts in a language both designers and business professionals will understand. Cole suggests other books and authors within this book as you gather more experience. In truth, this book provides a robust and accessible, allowing you to talk about visualisation in a language your audience always understand.
The storytellingwithdata.com provides additional context for those interested, and the monthly #SWDchallenge encourages members to hone their skills, challenge assumptions, and try new tools and methods. In addition to this book, my friend also gifted me the Storytelling with Data: Let’s Practice! Workbook.
Reference : NUSSBAUMER KNAFLIC, Cole. Storytelling with Data: A Data Visualization Guide for Business Professionals. John Wiley and Sons, New Jersey, 2015. ISBN 10: 1119002257
Appreciation: Christmas 2021, my friends and I shared our goals for the New Year. For the first time, I spoke loudly about my yearning to improve all my cross-sectional skills in Data. I jokingly said I wanted exam vouchers, textbooks and course vouchers as gifts this year.
Regardless, it was a big surprise when I received the notification that Storytelling with Data had been purchased in my name with a heartfelt message from my friend Tolu. Cheers to great friendships!