Making Smarter, More Persuasive Charts

Here are my takeaways from Scott Berinato book, Good Charts.

With the advancement and use of technology at every level, we have access to an abundance of data. We are faced with many decisions each day and if we have data available, it makes the decision-making process easier. We have more confidence in decisions that are data-driven.

Data visualization plays a key role in helping with decision-making. Simplicity is the key to data visualization. Visualization should communicate a single message.

There is no hard-and-fast rule to identify whether data visualization is good or not. But when making any chart, it helps to know the answer to most of the following questions to help decide how to visualize the data.

Who will see it?
What do they want?
What do they need?
What idea do I want to convey?
What could I show?
What should I show?
How will I show it?
What am I trying to show, say, learn, or prove?

If someone has to explain the chart, or visualization, that means that it is a poorly designed chart. If the chart can’t speak for itself, it has failed, like a joke whose punchline needs explanation. The charts should drive discussion but it should be about the ideas in the chart and not about the chart itself.

How do we see charts?

There are a few things we should be aware of when creating data visualization, such as how charts are looked at.

We do not go in order: Unlike text, we do not look at charts from top to bottom or left to right. Most of the time, headings are not even looked at first. So the chart should be very clear without even looking at the heading.
We see first what stands out: Make sure that what you are trying to convey stands out, e.g., with a bright colour
We see only a few things at once: A chart should not have more than five to ten variables. According to experts, we can’t distinguish more than eight colours.
We seek meaning and make connections: Our minds try to assign meaning to a visual and make connections between the elements presented, which sometimes are not accurate. That’s why visual elements presented together should be related in a meaningful way.
We rely on conventions and metaphors: There are some known conventions that are used by everyone, like green is good and red is bad, or time moves from left to right. If these conventions are not used, it can cause confusion.

Understand the why behind the chart:

Before creating any chart, decide the following:

Is the information conceptual (where the focus is on ideas and the goal is to simplify or teach) or data driven (where the focus is on statistics and the goal is to inform or enlighten)?

Am I declaring something (where the focus is on documenting or designing and the goal is to affirm) or exploring something (where the focus is prototyping and interacting and the goal is collaborating or discovering)?

Good Visualization

For making a good visualization, here are some guidelines:

Generate ideas by sketching the concepts for visualizing ideas Putting the data aside, as data limits how you think.
Use metaphors where applicable
Don’t overdesign.
Encourage brainstorming and capture many ideas
- Know what will be the heading of the chart
- Know the audience
- What setting will be used?
Make idea, not the design, pop
The charts should have consistent structure
- Title
- Subtitle
- Visual field (visual, axes, labels, captions, legend)
- Source line
Limit the eye travel on the charts. Place labels and legends to the close proximity to what they describe.
Like colours mean like items, so arrange the items accordingly
Colour saturation indicates a progression of values, e.g., from light to dark colours
Arranging and plotting the categories from one extreme to another, makes it more clear
Group the subcategories to bigger categories, while not losing the subcategory data. Use the same colour for each subcategory and label them.
The chart should be clear so that viewers are not stuck wondering what they are supposed to see.
- Make the purpose of the chart clear. The chart elements should be used to convey the idea instead of describing the chart structure. When chart elements are used to describe the chart structure, e.g., a title or subtitle repeating the axis labels, captions describing what the visual shows, it is a sign that the chart maker lacks confidence that the visual can convey the idea on its own.
- Add clarity by giving a title or subtitle to the question that visualization answers. Think about what I am trying to say or show?
Remove all ambiguities. Ambiguity in visualization generates stress to parse the data quickly and viewers tend to think more about the visual than the idea.
The general view is that simplicity indicates the absence of stuff, which is not always true. When making a chart, think about how little you can show but still convey the idea clearly. Remove redundant and unnecessary information.
To identify which elements should be kept, use the following flow:
- Is this element necessary?
  - Yes—is this element unique?
    - Yes, can this element be made simpler?
      - Yes, simplify and keep this element
  - No, consider eliminating
    - No -compare the similar elements and consider eliminating one of them
      - No, keep the element
Visualization is an abstraction and labelling everything is a concretization. If you feel that it is necessary to label everything, consider using tables. Use the purpose of visualization to decide what can be removed and what type of visualization should be used.
The simplicity of the chart decreases when we make it too colourful because we want it to be eye-catching or have lots of categories to spot. For each colour, ask yourself, Why do I need to make this distinction? Can it be combined with others as a group with a single colour?
Eradicate the deep rooted belief that more is better and complex is smart. Be confident in showing the information that is needed.
Use rich colours to make important ideas come forward and diminish other information with lighter or contrasting colours.

Persuade, not manipulate:

Do not manipulate the data in an effort to persuade. Ask the following questions and answer them honestly:

Does my chart change the idea or make it easier to see the idea?
If it is changing the idea, does it contradict the idea in a less persuasive chart?
If I am eliminating some data, does it hide something?
Would I feel betrayed if someone else presented a chart like this and I found out the changes I had made to the chart?

Presenting charts

Pause after displaying a chart, as the brain needs some processing time when it sees a visual. If you start talking right after showing the chart, most people will not understand the chart as they did not get time to process the information.
Do not read the chart, it will disengage the audience, as they have already looked at the chart and if the chart is clear, they understand what it is, so no explanation is needed. Presentation guru Nancy Duarte said, “Don’t project the idea that you are showing a chart. Project that you’re showing a reflection of human activity, of things we did to make a line go up or down. It’s not ‘here’s our Q3 financial results’, it’s ‘here’s where we missed our target’”
Only explain the chart, if it is unusual or complicated, before discussing ideas.
Reference charts can be used when you have to show the goal and compare it with the current situation. e.g., this is the desired form of the chart and this is how it is right now.
When you have something important to say, turn off the chart, as it is distracting and people mostly focus on graphs, so do not give the audience a choice about where to focus.

Engagement tips:

Withhold showing the whole picture at once. Ask questions first so the audience can speculate on what it will ultimately show.
Explain the value of the chart before going into detail.
Tell a story using the chart. Audiences reach understanding more quickly when stories are used.

Reference: Good Charts: The HBR Guide to Making Smarter, More Persuasive Data Visualizations by Scott Berinato

Making Smarter, More Persuasive Charts

Tayyaba Sharif