Choose the Best Data Visualization Format for Your eLearning

Data visualizations are an indispensable tool in eLearningand performance support. They communicate stories, help people explore data,and clarify complex information. The most attractive format is not always theright one for your data; being able to choose the best data visualization formatfor your eLearning can be challenging. Here’s some help.

Consider your goal

Not all datavisualizations in eLearning accomplish the same goal. Some are greatfor showing change over time or comparing two variables. Others show abigger-picture perspective or indicate trends. In 1984, WilliamCleveland and Robert McGill applied the science of visual perceptionto the creation of graphs. They came up with a hierarchy of encodingprecision—chart types—with chart types that allow humans to visually perceiveprecise information at the top, ranging down to less-precise types that excelat presenting an overview or trend view.

Most humans can easily perceive differences in line length,height, or position along an axis, making bar charts and scatter plots highlyaccurate ways to enable readers to compare values. Area, size, and slope aremore difficult for humans to perceive visually, so pie charts and bubble chartsare good for showing a trend or big picture but make comparing exact valuesdifficult. Which type to use and where to land on that scale depends upon thegoal of the visualization.

When comparing sales results, test scores, participationrates, or similar numerical data, whether among learners, between divisions, orover time, a chart type that offers accurate, precise perception is ideal,whereas a pie chart or bubble chart might not be the right format. When showingregional differences, though, a map with bubbles or colors representing greateror lesser values can convey the needed information; if readers need fine detailas well as the big picture, multiple charts will be needed to tell the completedata story.

Chunking content in data visualizations

A common error that designers make when creatingvisualizations is putting too much information into a single chart. Chunkingcontent—useful in presenting large amounts of complex information—works in datavisualizations as well as in text.

Imagine that you want to show the sales results of fiveregional teams, each with four sales representatives. You’ve got quarterly datafor the past five years. That’s a lot of information. You could create a singlevisualization with a line for each sales rep (20 lines), using a differentcolor to show each region or an additional line for each region’s average. Youmight want to create this one comprehensive chart—but it could be very hard fora learner to tease out the salient information from all those colored lines.

Instead, you might consider “sub-setting” the data—createthe comprehensive chart, but replicate it five times. Each one would emphasizeone team’s performance with a bold, striking color and show—but gray-out—theremaining data. This way, in five separate images, you offer more a focusedview of each team’s performance, showing the individual members’ results. Asixth chart could compare the averages of the five teams. This presents all ofthe data but does so in a way that offers manageable chunks for readers tostudy and understand.

Choose the right chart—or charts

Before designing a data visualization, decide whether thereaders should be able to:

• Compare values
• Identify averages and outliers
• Figure out proportions or relative parts of awhole
• Analyze trends
• Understand the relationships between two (ormore) variables
• See changes over time
• See rank or hierarchy
• Compare values with a fixed reference point
• Understand flow or movement between conditions
• Understand spatial relationships or geographicpatterns

Several free online tools show a variety of data visualizationformats that can accomplish each of these goals.

• The Data VisualizationCatalogue offers a searchby function or a list of visualizationtypes. Each one includes an illustration and an explanation.
• The Financial Times graphics departmentcreated an enormous poster, titled “Visual vocabulary,” that explains variouschart types and their purpose. It’s availableonline and for download.
• Ann K. Emery, a data scientist and teacher,created an interactive online catalog of what she terms essential charttypes. Each has a brief explanation, an illustration, and examples. Thesecan be sorted by purpose.

With practice and experience, it becomes easier to choose the best datavisualization format for your eLearning. The next step is learning to createthem! Start by mastering three basic data visualizations that are likely to be useful over and over in eLearning, and learn moreabout how to use data and analytics in eLearning at The eLearning Guild’s Data& Analytics Summit, August 22 & 23, 2018.