I’m currently working on a new analytics product and want to learn more about data visualization and communication. Several of my friends working in data recommended me this book Storytelling with Data. The author Cole Knaflic has a BS in Applied Math and an MBA. She worked as a manager on the Google People Analytics team. She has a mixed background of mathematics, business, and design. She also has a website updating new principles, workshops, and exercises about data analysis if you are interested: http://www.storytellingwithdata.com/.
I enjoyed reading the well structured book and the sleek touch of the page texture. However, many principles were quite repetitive especially towards the last chapters of the book. However, the main benefits of reading the whole book are that Cole provided graphs before and after the design and she walked readers through some case studies, so as a reader, I can understand how would it look like implementing these principles in the real world.
I summarized the most insightful points here that I found from reading it. This article is perfect for someone 1) who wants to read this book but doesn’t have time or; 2) who is curious and not sure if he/she wants to read it yet or; 3) who has a data visualization and communication project and wants to have a sanity check. Because it is a summary, I don’t provide details of some of the principles, but please Google them directly if you are curious about their exact meanings.
Context
Who
- Who is the decision maker?
- What’s your relationship with the audience?
What
- Why they should care?
- Check Big Idea by Nancy Duarte — so-what in a single sentence
- What is the desired outcome?
How
- Data is the supporting evidence of the story you will tell
- Use tactics: 3-minute story and storyboarding (use Post-it to rearrange)
Charts
Simple text
- When reducing from multiple numbers down to a single one, think about what context may be lost
Tables
- Great to a mixed audience who will each look for their particular row of interest
- Not a good idea for live presentation
Heatmap
- Always include a legend
Slopegraph
- Show relative increases and decreases or differences across various categories between the two data points
Bars
- Go-to graph type for categorical data
- Must have a zero baseline
Waterfall chart
- Pull apart the pieces of a stacked bar chart to focus on one at a time
- To show increases and decreases, and the resulting ending point
Stacked horizontal bar chart
- Visualize portions of a whole on a scale, for example: survey data collected along a Likert scale
Area
- Visualize numbers of vastly different magnitudes
To avoid
- Pie chart (replace with a bar chart; *other opinions in data community say, however, pie chart can be used if for just few categories and can clearly represent the percentage)
- 3D
- Secondary y-axis
Visual Principles
Avoid clutter
- Avoid cognitive loading, processing that takes up mental resources but doesn’t help the audience understand the information
- Maximize data-ink by Edward Tufte
Visual order
- People usually start reading from upper left of a chart
- If there is only one thing that is really important, making that the only thing on the page
Use of contrast
- The more things we make different, the lesser the degree to which any of them stand out
Preattentive attributes
- Help direct your audience’s attention to where you want them to focus it
- Create a visual hierarchy of elements to lead your audience through the information you want to communicate in they way you want them to process it
- People can keep about four chunks of visual information in their short-term memory at a given time
- Use preattentive attributes for explanatory story analysis, not for exploratory analysis
Color
- Should always be an intentional decision; identify one or two brand-appropriate colors to use as your “audience-look-here” cues
- Use varying color saturation of a single color (a heatmap)
- Design with colorblind in mind — 8% of men and half a percent of women are colorblind
Highlight
- Bolding is generally preferred over italics and underlining
Storytelling
Persuasion
- Unite an idea with an emotion, arousing the audience’s attention and energy
- McKee “subjective expectation meets cruel reality” — an event that throws things out of balance
- Don’t communicate for yourself — communicate for your audience. What motivates your audience?
Structure
- Bing (intro)
- Bang (content)
- Bongo (conclusion): call to action; repeat (“here is what we covered”)
Tactics
- Have action titles (not descriptive ones)
- Get feedback from your friends and relatives — fresh eyes
