Improving Data Visualization Using Five Principles of Gestalt

Michael Cleavinger, Director of Space Transformation, PepsiCo [NYSE:PEP]
100
159
30

Michael Cleavinger, Director of Space Transformation, PepsiCo [NYSE:PEP]

As data visualizations become ubiquitous in the workplace they can easily lose their salience. I propose five ways to improve your data visualizations using the principles of Gestalt, which is the whole being greater than the sum of the parts:

1. Understand the Whole
2. Find Your Story
3. Use the Laws of Gestalt
4. Make it Interactive
5. Apply it Yourself

1. Understand the Whole

Have you ever looked at the clouds and saw an animal or figures in the stars? The mind longs to understand what it perceives and is adept at creating shortcuts to quickly and accurately comprehend our world. You can use these same tricks and shortcuts to enhance data visualizations, to improve understanding, and bring clarity to your message. This is Gestalt: focusing on the whole rather than the individual parts.

2. Find Your Story

You must understand what you’re attempting to share. What is your ‘whole?’ What story are you trying to tell? What parts of the story are necessary to build a comprehensive whole? What would be lost by not knowing all those parts?

“Today’s graphics use multiple shapes and colors to cover up a paucity of data.”
                                                                                                                              -
Edward Tufte

Data visualizations are not an afterthought to fill white space. What is the one point you want the consumer to take away? Using this guide, analyze the most important pieces of your story, what should be emphasized, what is the desired impact, and what potential issues exist. You are telling a story and must know your information’s ‘characters’ to develop, engage and inform.

  â€‹Your visualization should put the focus on your story, not the visualization   

3. Use the Laws of Gestalt

The human brain is adept at taking bits of disparate information and instantaneously creating concise and meaningful context for understanding. We constantly perform this subliminal work to understand the world. Data, as a rule, is meaningless without context. Let’s focus on three laws of Gestalt to derive context.

• Law of Proximity – We innately perceive items placed close together as a group. In the picture below, we perceive four distinct groups with one group being larger and more distinct than the others. Only after consideration or prompting do we realize there are 72 individual circles.

• Law of Similarity – We naturally group items by similar attributes such as color, shape, or size. Below we see three rows of black dots grouped together that draw our focus.

• Law of Closure – We perceive objects to be whole even when not complete. Below we can easily identify the circle and rectangle despite the missing lines or arcs. Likewise, we create figures in the sky by connecting the stars.

Businesses capitalize on these laws to provide context in their logos. For example: In the middle of the Tostitos logo there are two people dipping chips in salsa (Law of Closure and Law of Proximity); Fed Ex has a hidden arrow between ‘FedE’ and ‘x’ (Law of Closure); Baskin Robbins’ initials and 31 flavors are blended together (Law of Similarity). Using these laws draws the consumer in by conveying information and providing additional context quickly and effortlessly.

Consider and use the laws of Gestalt to define how you share the story. Ask yourself why you are adding color, what is the purpose for changing an item’s shape, whether the visualization provides clarity, and the value of using a pie chart. Analyze the need for each point, reconsidering your story. Except for the question on the pie chart, don’t use pie charts.

4. Make it Interactive

People actively seek information and want to interact with it. Interactive data visualizations provide an optimal way to convey a story as a whole. Don’t confuse interactive data visualizations with reporting dashboards, they are different and have diverse purposes. Interactive data visualizations allow for conversations about the story. By moving to interactive visualizations you will benefit by discussing the story, not the parts individually. It also allows for clarifying the parts to help others see the whole.

5. Apply it Yourself

“The greatest value of a picture is when it forces us to notice what we never expected to see.”
                                                                                                                                                  -
John W. Tukey

Use visualizations from the beginning of your analysis to the end. Build visualizations using the Gestalt principles for analysis, not just the final presentation. Implementation throughout the process will help you develop and improve your story. Experiment! Change colors, sizes, and shapes, zoom in, zoom out, flip the axis of a grid and change the type of visualization. Add new data you haven’t previously put together before, print it out instead of looking at the screen. Ask a colleague to verbally explain it to you. Force yourself to look at the information in different ways to find new insights or a better way to share the information.

Your visualization should put the focus on your story, not the visualization. You need to know what you’re sharing to know what to build and what tool to use. Sometimes you need a hammer, sometimes you need a shovel.

Find a tool that enhances your storytelling ability. There are several effective visualization tools and packages: D3, Plotly, Leaflet, Tableau, Spotfire, Power BI, Alteryx, Qlik, Splunk, SAS Visual Analytics, and many more. Don’t let the tool dictate your story. Build your story with interactive visualizations, share the visualizations, present the visualizations. Don’t just walk someone through a PowerPoint; take them on a journey.

As you change the way you approach visualizations, use reference books from the great Edward Tufte, a pioneer of data visualization, or Nathan Yau who is making data visualizations and interactive visualizations more accessible. Continually look for new ideas from websites like FlowingData, FiveThirtyEight, GraphIQ, WolframAlpha. And above all else, focus on your story and how your data visualizations make your story more complete, more relatable, and provide additional context in a simple and quick manner.

Read Also

Fostering the Innovative Culture through Proof of Concept Projects

Karl Hightower, VP of Transformational Business Solutions, Rent-A-Center [NASDAQ:RCII]

Improving Predictive Analytics with Data Visualization

Fadi Elawar, Technical Consultants Team Lead, iDashboards

Simplify and Strengthen Your Approach to Data Visualization

Rolf Olsen, Chief Data Officer, Mindshare

Balancing Analytic and Visual Creativity for Effective Data Visualizations

John Lucker, Risk and Financial Advisory Principal and Global Advanced Analytics Market Leader, Deloitte & Touche LLP