There are six elements that are important for good storytelling:
- Understand the context
- Choose an appropriate visual display
- Eliminate clutter
- Focus attention where you want it
- Think like a designer
- Tell a story
Understanding situational context including the audience, communication mechanism and desired tone, is important. Data visualization and communicating with data in general sits at the intersection of science and art.
We have exploratory and explanatory analysis. Exploratory is about understanding data. But in visualization we are all about explanatory.
Before starting we should know:
- To whom we are communicating?
- What do we want our audience to know or do?
- How can we use data to help make our point?
Who: audience and us.
Narrow your audience, you cannot talk to too many different people with disparate needs.
What: action, mechanism and tone.
You should always aim for your audience to know or do something. If you cannot propose action directly, encourage discussion toward one. Mechanism is about how will we communicate to our audience. Do not use slides as teleprompter. They can remind you of next topic, but they should not act as speaking notes. Avoid slideument – when both slide decks and documents are meant as one, single document.
What data are available.
While it sounds easy, being concise is often more challenging than being verbose. Blaise Pascal: “I would have written a shorter letter, but I didn’t have the time.”
The 3-minute story can help with being concise. The Big Idea boils the do-what down even further: to a single sentence. Nancy Durante discusses this concept in her book Resonate. Big Idea has three components:
- Unique point of view.
- What’s at stake.
- Complete sentence.
Storyboarding creates structure of your presentation. Use a whiteboard.
Simple text – when you have just a number or two to share, simple text can be a great way to communicate. The fact that you have some numbers does not mean that you need a graph. Any time you reduce from multiple numbers down to a single one – think about what context may be lost in doing so.
Tables – interact with our verbal system, which means that we read them. Tables are great communicating to a mixed audience whose members will each look for their particular row of interest. It is also easier to communicate multiple different units of measure. Using a table in live presentation is rarely a good idea. Borders should be used to improve the legibility of your table. The data should be what stands out, not the borders.
Heatmap – we can use color saturation to provide visual clues. You should use legend to help reader interpret the data.
Graphs – interact with our visual system. They fall into four categories: points, lines, bars and area. Chart is the broader category, including graphs, but also maps and diagrams.
Points – scatterplot can be useful for showing the relationship between two things.
Lines – are most commonly used to plot continuous data. Often is in some unit of time: days, months, years. We have standard line graph and slope graph. The line graph can show a single series of data, two series or multiple. Slope graph is best used when you have two time periods or points of comparison and you want to quickly show increase or decrease of difference.
Bars – are sometimes avoided because they are common. They are easy for our eyes to read. It is important to always have a zero baseline, otherwise you get a false visual comparison. Only with line graph you can get away with no zero baseline. When graphing data, a common decision to make is whether to preserve the axis labels or eliminate the axis and instead label the data points directly. In general, the bars should be wider than the white space between bars.
The plan vanilla bar chart is the vertical bar chart. Stacked vertical bar charts are meant to allow you to compare totals across categories and also see the subcomponent pieces within a given category. They can be absolute numbers or percentage. The waterfall chart can be used to pull apart the pieces of a stacked bar chart to focus on one at a time, or to show a starting point, increase and decrease, and the resulting end point. Horizontal bar chart is go-to category, since they are really easy to read. It is especially useful if category names are long. Always be thoughtful about how categories are ordered.
Area graphs – avoid at all costs, except when you need to visualize numbers of vastly different magnitudes.
Infographic is a term that is frequently misused. An infographic is simply a graphical representation of information of data.
You really should avoid: pie charts, donut charts, 3D and secondary y-axis. Pie charts are hard for people to read. With pies, we are asking our audience to compare angles and areas. With donut chart, we are asking our audience to compare one arc length to another arc length.
We experience cognitive load anytime we take in information. Cognitive load can be thought as the mental effort that’s required to learn new information.
Edward Tufte refers to the data-ink ratio – the larger the share of a graphic’s ink devoted to data, the better (other relevant matters being equal). Nancy Durante in her Resonate refers to maximizing the signal-to-noise ratio. Signal is the information we want to communicate.
Clutter are visual elements that take up space but don’t increase understanding.
Gestalt Principles of Visual Perception has six principles:
Proximity – we tend to think of objects that are physically close together as belonging to part of a group. Same goes for similarity and enclosure. The closure concept says that people like things to be simple and to fit in the construct that are already in our heads. When looking at objects, our eyes seek the smoothest path and naturally create continuity. We tend to think of objects that are physically connected as part of a group.
When design is thoughtful, it fades into the background so that your audience doesn’t even notice it. When it’s not, however, your audience feels the burden.
Some elements that can help with proper design. Alignment. Usually, it should be to the left. Since we are readers from the left. Without other visual cues, your audience will typically start at the top left on the page or screen and will move their eyes in a “z” shape. Diagonal elements like lines and text should be avoided. Reading rotated text for 45 degrees is 52 % slower. White space in visual communication is as important as pauses in public speaking. Clear contrast can be a signal to our audience, helping them understand where to focus their attention.
- Remove chart border
- Remove gridlines
- Remove data markers
- Clean up axis labels
- Label data directly
- Leverage consistent color
Preattentive attributes are: size, color and position on the page. They can be used to draw attention where it is needed or create a visual hierarchy of elements.
People look with eyes, but see with brains. We have three types of memory: iconic, short-term and long-term. Iconic is super-fast. Short-term has limitations. People can keep four chunks of visual information in their short-term memory at a given time. Long-term is built up over a lifetime. It is important for pattern recognition and general cognitive processing.
Verbal memory is accessed by a neural net, where the path becomes important for being able to recognize or recall. Visual memory, on the other hand, functions with specialized structures. Our brains are hardwired to quickly pick up differences we see in our environment.
Elements to be used to encode quantitative information:
- Spatial position
- Line width
Preattentive attributes in text:
- Separate spatially
- Outline (enclosure)
- Underline (added marks)
Studies have shown that we have about 3-8 seconds with our audience, during which time they decide whether to continue to look at what we’ve put in front of them or direct their attention to something else.
Size matters. Relative size denotes relative importance.
When it comes to the use of color: use it sparingly, use it consistently, design with the color-blind in mind, be thoughtful of the tone color conveys and consider whether to leverage brand colors. Grey is great base color; blue is good for attention-grabbing color. When we use too many colors together, beyond entering rainbow land, we lose their preattentive value. A change in color signals just that – a change. Leverage the same schematic throughout the communication. Due to colorblind people (8% of men and 0,5 % of women) avoid using shades of green and red together.
Position is also important. Most of audience will start at the top. If something is important, eliminate the work of audience to find it, put it at the top.
Form follows function. In the field of design, experts speak of objects having “affordances”.
- Highlight the important stuff
- Eliminate distractions
- Create a clear hierarchy of information
Don’t highlight more than 10%. Bold is generally preferred over italics and underlining. Avoid using different fonts as a highlight technique. Color should be used sparingly and potentially in combination with other elements like bold. Inversing elements should also be used sparingly.
Eliminating distractions. Antoine de Saint-Exupery: “You know you’ve achieved perfection, not when you have nothing more to add, but when you have nothing to take away.”
Some actions to eliminate distractions:
- Not all data are equally important
- When detail isn’t needed, summarize
- Ask yourself: would eliminating this change anything
- Push necessary, but not-message-impacting items to the background
Two specific strategies related to accessibility in communicating with data: don’t overcomplicate and text is your friend.
- Make it legible
- Keep it clean
- Use straightforward language
- Remove unnecessary complexity
Text plays a number of roles in communicating with data:
- Tell a story
Thoughtful use of color, alignment and white space are components of the design that you don’t even notice when they are done well.
Some strategies you can leverage for gaining acceptance in the design of your data visualization:
- Articulate the benefits of the new or different approach
- Show the side-by-side (if new approach is really superior)
- Provide multiple options and seek input
- Get a vocal member of your audience on board
When we have actual and forecast data, we can use visual cues to draw a distinction. Red can be used as attention-grabbing color. But maybe tone it down a little. Words make the visual accessible.
Story is a time-tested structure: humans have been communicating with stories throughout history. Conflicts and tension are an integral part of story. Stories unite an idea with an emotion, arousing the audience’s attention and energy. At a fundamental level, a story expresses how and why life changes. Stories start with balance. Then something happens. Subjective expectation meets cruel reality.
Kurt Vonnegut in his How to write with Style suggests:
- Find a subject you care about
- Do not ramble, though
- Keep it simple
- Have the guts to cut
- Sound like yourself
- Say what you meant to say
- Pity the readers
The essential elements of story: setting, main character, unresolved state of affairs and desired outcome. End with a call to action. The most beautiful data visualization runs the risk of falling flat without a compelling narrative to go with it. A strong narrative can overcome less-than-ideal visuals.
Horizontal logic. Audience can read just the slide title of each slide throughout slide deck. It is important to have action titles (not descriptive titles) for this to work well. Vertical logic means that all information on a given slide is self-reinforcing. We also can use reverse storyboarding and a fresh perspective.
Light elements on a dark background can create a stronger contrast but are generally harder to read. There should be logic in the order in which you display information. A spaghetti graph is a line graph where the lines overlap a lot, making it difficult to focus on a single series at a time. Don’t do it. Emphasize one line at a time. And we can also pull lines apart horizontally and vertically. As alternatives to pie charts, we can: show numbers directly, use simple bar graph, use stacked bar graph or slope graph.
 In the book on page 30
 In the book on page 132