How to Get the Most Out of Data Visualization

Data visualization is the term applied to the graphical representation of data. By creating visual elements, such as charts, graphs, and even animation, data visualization tools offer an easily accessible way to see and understand data trends and patterns.

The best way to get value out of data visualization is to know your audience, advises Christopher Goranson, distinguished service professor at Carnegie Mellon University’s Heinz College of Information Systems and Public Policy.

Data visualizations are frequently built by individuals intimately familiar with the data. “But, often, the person who needs to know the information is much less familiar with it, if at all,” Goranson says. He believes that it’s always a good idea to ask for feedback on early designs, and to remember that it’s the designer’s job to ensure that the data visualization makes sense.

Ensuring that the data is accurate, complete, and at the appropriate level is necessary to ensure great data visualization, says Michael Willock, CTO, investments and CIO, corporate function, at Liberty Mutual Group. “It’s much less about a tool or specific technology, and more about understanding your customer.”

Insights can be gained or lost by how well users’ needs and expectations are met. “The creation of great data visualizations starts long before the creation of a dashboard, report, chart, billboard, or ticker,” Willock says. “With a deep understanding of your customer, and by partnering with them to research, design, test, iterate, and build visualization solutions, you’ll ensure value is delivered.”

Related:Guinness, an Aircraft Carrier, and the History of Data

It’s important to fully understand the user’s needs, says Fernando Cuadra, a consultant at technology research and advisory firm ISG. “A financial adviser will have different needs than a product management team looking for customer behavior data,” he explains. “The best way to learn the true needs of your audience, and to [help them] get the most value out of your data, is to run a design sprint specifically around your target audience,” Cuadra suggests. “This technique applies design thinking practices into what is usually an engineering process.”

Data Visualization Attributes

Data visualization is too broad an area for a single tool to satisfy all user needs, Cuadra says. He notes that IT teams should gauge the target user’s data visualization maturity before developing and deploying an offering. “A complex tool with great attributes will never be used in a low maturity environment,” Cuadra warns. In contrast, basic tools will frustrate a highly mature data audience, often to the point where users will be tempted to develop their own data analysis tools.

Related:Dispelling 5 Misconceptions About Data

Regardless of the target audience, a data visualization solution should always be purposeful, targeted, and easy to consume, Willock suggests. “By using data visualizations, you can create deeper connections with customers, reduce total cost of ownership over time, and create a competitive advantage,” he explains. Data visualization benefits include reduced time to insight and limited cognitive load, Willock adds.

Most business users seek simple, straightforward data visualizations. “For advanced, highly technical use cases, such as scientific research, rich graphical techniques are very important,” Willock advises, adding, “it can be difficult for tools to achieve both goals.”

Tools, such as Tableau, can help enforce good design practices out of the box since many default settings are designed to assist user comprehension, Goranson says. On the other hand, “tools that pick random colors or recommend stylistic choices can get in the way and are less helpful,” he notes. “For example, a tool that recommends a lot of 3D graph choices or unnecessary treatments, like drop shadows, shading, or cartoonish renderings.”

Common Mistakes

Over-engineering reports and dashboards is a common mistake, Cuadra says. Developers should aim for simplicity. Data visualization should be viewed as a story-telling exercise that presents answers to important questions in a clear and succinct way. “More isn’t necessarily better in data visualization,” he cautions. “Additionally, if the data doesn’t clearly state what the next action should be, then we’ve failed in our data visualization journey.”

Rushing to deliver data without understanding the use case or customer need is a common pitfall, Willock says. “This often leads to grids of extraneous data with a multitude of filters or, at the other extreme, fancy visuals that don’t answer the important questions people are looking to get answered.”

“Beware!” Cuadra warns. “Data can tell you the story you want to hear, not the data you need to know.” This fact becomes most evident in executive environments, he notes. “Question everything when it deals with data.”