Data visualizations are an easy-to-use, essential tool for understanding and communicating the complex stories that data and analytics can tell. Creating multiple visualizations allows a developer to explore the data and learn some of the interesting ideas, patterns, and surprises it contains. And, using data visualizations in eLearning provides a way to enhance learner comprehension and engagement while revealing the many stories hidden in otherwise incomprehensible data tables.
Data visualization enables exploration
A dataset usually consists of numbers—hundreds, even thousands of numerical values that can describe virtually anything. Datasets are often presented in tables or spreadsheets with myriad lines and columns.
Looking at these numbers can uncover information—it is technically feasible to identify an outlier or a high value, for example, or to calculate an average using only a data table. But our limited human brains lack the ability to really grasp all of the information contained in a dataset simply by looking at the sea of numbers.
“Tables alone are definitely not sufficient to give us an overview of a dataset,” Gregor Aisch wrote in the Data Journalism Handbook, a free online resource.
Some analysts will look at a few descriptive statistics for a dataset and believe that they can pull the important information out of those values, which could include mean, variance, and correlation between x and y. But, as Anscombe’s quartet illustrates, datasets with the same descriptive statistics can tell very different stories—differences that are obvious when the datasets are graphed or turned into visualizations.
Data visualizations enable readers to explore data and see patterns or identify trends and outliers. Perhaps most importantly, a data visualization enables exploration of a dataset by raising questions. Looking at the visualizations can, for example, trigger questions about why some data values are so different from the rest of the set (outliers) or what might happen if two or more of the variables were combined—or even why the numbers are what they are.
Data journalist Alberto Cairo, who holds the Knight Chair in Visual Journalism at the School of Communication of the University of Miami, emphasizes that data exploration does not provide answers or explanations for the numbers; it offers a starting point. The analyst or developer must confirm the data by talking with experts and digging deeper into the trends, patterns, and information in the data.
Data visualizations facilitate storytelling
In eLearning and performance support, a key role for data visualizations is telling stories—putting the information that the data capture into a context and form that learners can understand and apply on the job.
Cairo cites four principles that data visualizations must follow in order to successfully communicate or tell a story:
- Any good graphic—or act of communication—begins with good data.
- A successful data visualization attracts and engages readers’ attention.
- Successful data visualizations do not frustrate readers; the graphic must strike a balance between an engaging or surprising appearance and a clear, easily understood presentation of information.
- Communicating data does not mean over-simplifying it. At the same time, it’s important not to overload readers.
A successful visualization has the right amount of data needed to tell the story fully and accurately. Too much information can wreak havoc with reader comprehension. Including too little data risks omitting key details and communicating a misleading or even a false story. “Data visualization clarifies information; it doesn’t simplify,” Cairo said in an online lecture.
“Data visualization is only successful to the degree that it encodes information in a manner that our eyes can discern and our brains can understand. Getting this right is much more a science than an art, which we can only achieve by studying human perception. The goal is to translate abstract information into visual representations that can be easily, efficiently, accurately, and meaningfully decoded,” Stephen Few wrote in “Data Visualization for Human Perception.”
Creating the data visualization that will accomplish these goals requires enough understanding of the data to figure out which format—or formats—the data visualizations should take, choosing the correct tool, and adding a title and annotations that provide any needed context.