Data visualization tools need to be intuitive

As more business end-users get access to analytics tools, the way data is presented will need to enable "cognitive" visualization in order for them to better make sense of the insights.

The growing volume and complexity of data that companies are collating and analyzing, as well as the empowering of more business end-users to access such insights, have raised the importance of intuitive data visualization interfaces for analytics tools, say industry watchers.

John Brand, vice president and principal analyst at Forrester Research, said there is now more general users without technical knowhow handling data analytics than in the past, when such tools were limited to the tech department.

At the same time, the relationship between different data sets has gotten more complex, Brand noted. Traditional data visualization approaches were simplistic in representing the correlation such as through rows and columns on Excel spreadsheets.

Today, though, there is a greater emphasis on integrating data from a wide variety of sources, so new methods of visualizations such as infographics, interactive bubble charts and 3D landscapes are increasingly needed, he pointed out.

The idea is to enable "cognitive" visualization in that the user is able to view different sets of data based on a condition, event or scenario as opposed to a chart, graph or table, the analyst explained.

Raffael Marty, founder and CEO of PixlCloud, a big data visualization app developer, added the concept of data visualization is not new to the industry since it has been used in business intelligence (BI) software, but whether existing interfaces helped end-users better understand the information is debatable.

These tools have never really catered to "real requirements" from the start, Marty said. Instead of figuring out what people want in terms of user experience before designing the tool, IT departments will wait until end-users pose a problem they'd like solved before they build a program and "add some user experience pixie dust" to it.

By doing so, most companies have to rely on data scientists who know how to use the program in order to extract actionable information, he added.

"We still don't have intuitive interfaces. It's a really hard problem of balancing technical capabilities with features that a user really needs," he stated, adding a good data visualization program means users should not need to learn how to use it.

"We're just starting to understand what end-users need, and how the interfaces have to look in order for them to be efficient. But currently, it takes technical people who know the business end-user side to engineer the tools," the CEO pointed out.

Bhavish Sood, research director at Gartner, said data visualization tools need to be more interactive, going beyond that of lines, bars and pie charts.

Interactive visualization technology gives users the chance to actively explore data visually and that helps them understand and assimilate data more effectively than through rows, columns, figures and static charts, Sood pointed out.

Not replacing traditional methods
Forrester's Brand noted the new methods of data visualization will not replace traditional methods, which are still relevant. Rather, companies should have the choice to pick the appropriate tool and strategy based on their desired objectives and outcomes, he said.

For instance, a salesperson might find charts the most effective way of presenting the company's data to clients, whereas a service representative might feel a semantic map is easier to sieve out the context of the most-quoted words used in customer complaints, the analyst elaborated.

He added there is no correct answer to what constitutes the most effective visualization tool for each individual's needs.

"Who is the arbiter of good taste when it comes to data visualization? Many say it's the creative design teams more than the IT folk. Others say it's the marketing or the BI specialists. The only thing that matters is the organization has a robust process for identifying the most likely candidate and applying the best tool for the job," Brand said.