[Skip to Content]
Visit us on Facebook Visit us on FacebookVisit us on Twitter Visit us on TwitterVisit our RSS Feed View our RSS Feed
NEC Spotlight November 14th, 2024
CategoriesCategoriesCategories Contact UsContact Us ArchivesArchives Region/OfficeNEC Main Site SearchSearch

Jan

29

Date prong graphic

The Zen Trend in Data Visualization

Posted by on January 29th, 2016 Posted in: Data Visualization


You may have noticed there is a data visualization bandwagon out there and the OERC bloggers are on it. We like to write about tools and resources that can help you communicate visually.  There’s a bit of self-interest here. We want to fuel this bandwagon because we personally prefer to review a one-page visual synopsis of data over a table-dense written report.

Unless, that is, we have to grapple with a poorly designed data visualization. These are displays with too many details; unnecessary icons; many variables piled into one chart. It can make your head spin.

That isn’t to say you have to be an artist to do good visual displays. Quite the contrary.  Most information designers adamantly state that data visualization is about communication, not art. However, you have to know how to design with a purpose.

To understand the basics of good design, you need to understand why humans respond so well to visual displays of data.  That is the topic of an excellent blog article by Stephen Few, a thought leader in the data visualization field. First, he advocates that data visualizations aid users in performing three primary functions: exploring, making sense of, and communicating data.  Evaluators would add that our ultimate goal is to help our users apply data in planning and decision-making. To that end, Few argues that data visualizations should be designed to support readers’ ability to perform these four cognitive tasks:

  1. See the big picture in the data
  2. Compare values
  3. See patterns among values
  4. Compare patterns

To demonstrate Few’s point, I am sharing a sample of the OERC Blog’s monthly site statistics.  The blog dashboard uses a table format to display total views per month. The table shows monthly view counts starting in June 2014, when we started tracking site statistics.

Table showing total page views per month for the OERC blog from June 2016 to Jan 2016. The numbers generally get higher over time, but there is fluctuation month to month. It is difficult to see trendts

 

Line graph of data shown in previous table (June 2014 to January 2015) It shows total page-view-per-months. The general upward trend is much more apparent than in the table. 2015 has more page views than 2014 for all months. There are some times during the year where there is a zigzag pattern, with high viewership one month followed by lower readership the next. This would suggests a need to look at topics or schedules of the blog readership.

Now here’s the same information presented in a line graph created in Excel.  You quickly see the big picture: Our page views are trending upward.  You can easily compare the direction of month-to-month traffic.  The chart allows comparison of monthly patterns in 2014 and 2015.  It’s the same data, but its almost impossible to see any of these findings in tabled data.

Dataviz design experts like Stephen Few are avowed minimalists. They hate chart junk, such as gridlines and data labels.  They have an affinity for small multiples, which are series of graphs displaying different slices of data.  (If you have never seen small multiples, here’s a post from Juice Analytics with good examples) In general, they do not include any element that will hinder users’ ability to make comparisons, find patterns, and identify pattern abnormalities that may be indicators of important events.  Needless to say, most data visualization luminaries are not big on decorative features like gas gauges and paper doll icons. These, too, are viewed as unnecessary distraction.

Yet, there is a lot of chart art out there and you may wonder why. There is a distinction between data visualizations and infographics.  Alberto Cairo, Knight Chair in Visual Journalism at the University of Miami’s School of Communication,  wrote that data visualizations are tools for interactive data exploration while infographics are visual displays that make a specific point.  I tend to think of data visualizations as having users, while infographics have readers. Chart art may be more legitimate in infographics because it supports the primary message or story.

Yet, Cairo admits that the boundary between infographic and data visualizations is fuzzy. He noted a trend toward infographics with two layers: a presentation layer, and an exploration one.  The infographics have an obvious primary message, but readers are also presented with opportunities to explore their own questions. One of my favorites from the Washington Post is an example of an infographic with both layers.

That said, Cairo still argues that data design principles hold true for both data visualizations and infographics.  There is no excuse to drown your readers in images or to distract them with bling.

Zen is in; bells and whistles are out. The good news is that simple data visualizations do not require sophisticated software or design skills.  That’s not to say that simple is the same as easy. Good data visualizations and infographics take a lot of thought. For the more interactive data visualizations, you must identify how your users will use your data and design accordingly. For infographics, you need first to clearly identify your central message and then be sure that every element has a supporting role.

Want to start developing good visual design habits? I recommend Presenting Data Effectively by Stephanie Evergreen (Sage, 2013).

 

 

Image of the author ABOUT nnlmneo


Email author View all posts by
This project is funded by the National Library of Medicine (NLM), National Institutes of Health (NIH) under cooperative agreement number UG4LM012343 with the University of Washington.

NNLM and NETWORK OF THE NATIONAL LIBRARY OF MEDICINE are service marks of the US Department of Health and Human Services | Copyright | HHS Vulnerability Disclosure | Download PDF Reader