Ever since the first pie chart appeared in a presentation, organizations have used data visualization techniques to make colorless numbers more engaging and easier to grasp. And with the burgeoning amounts of data being captured, processed and analyzed by companies in the big data era, sophisticated visualization tools are being deployed to help parse everything from budgetary projections to customer feedback for business users. But installing the software is just the tip of the iceberg when it comes to using the technology to make data more understandable.
According to a report on advanced data visualization software published by Cambridge, Mass., analyst group Forrester Research Inc. in July, visualization technology has become a critical element of business intelligence (BI) and analytics programs. "There is too much data out there for [business users] to understand what the patterns are just by looking at numbers," said Boris Evelson, a Forrester analyst who co-authored the Forrester Wave report. Done effectively, visualization enables key data points from large sets of information to be displayed on a single screen in a way that greatly simplifies the analysis process, he said.
Evelson added, though, that putting data visualization tools in place is only the beginning of the process. "Before you can visualize data, you have to find it, extract it and integrate it," he said. Then, BI teams need to cleanse and model the accumulated data to ensure that the visualization-enabled reports and dashboards they build will provide accurate information. If you're expecting data visualization to help make "all [your] analytics problems go away, you would still have to do the legwork before you could get to the point where you'd start visualizing data," Evelson said.
In addition, BI teams need to have a broad knowledge of their organization's data and how it's used in order to deliver the right kinds of visualization capabilities to business executives and users, said David Stodder, director of BI research at The Data Warehousing Institute in Renton, Wash. "It's easy to get lost in the visuals and lose track of what this is really trying to convey in terms of information," he said.
TMI could hamper decision making
Another consideration is how much data to present to different users. "A lot of dashboards have way too much information," Stodder said. "Often, top-level executives are sitting there looking at a roomful of screens and looking at real-time data." Visualization technology can support that level of data immersion, he added, but users who receive so much information might stumble over it when making decisions -- it could simply be too much for them to process, according to Stodder.
Open door policy on data visualization?
By making information easier to understand, data visualization tools create the potential to spread business intelligence data more broadly in organizations. Executives at some companies might balk at the idea, wanting to keep access to BI data on a strict need-to-know basis. But Lee Feinberg, president and founder of DecisionViz, a consultancy in Westfield, N.J., advocates a more open approach.
Giving data access and visualization capabilities to larger numbers of business users can foster increased levels of internal collaboration, according to Feinberg -- and that, he believes, will result in tangible business benefits. "If more people in the organization can touch the data, work with the data and analyze the data, you can create much more value than if the data just sits in the database," he said. "You can increase the value of the company by letting more people have access to the data."
Source: Christine Parizo
The biggest mistake an organization can make on a data visualization project is to focus on how the available data can be visualized without putting the effort into a business context, said Lee Feinberg, president and founder of consultancy DecisionViz in Westfield, N.J. Instead of designing charts based on some data that's at hand, "start with the decisions [you're] trying to make with the data," he advised.
That approach pulls together the BI team and business units in an organization, Feinberg said: "They have to work more closely, because now the business has to be more explicit about what [they] need to drive those decisions and the BI team needs to make sure they can actually get that information in the appropriate form."
But once organizations get a firm grasp on the data that business users need to see, the design side of creating data visualizations comes into play. In their Forrester Wave report, Evelson and co-author Noel Yuhanna wrote that effective data visualization requires good visual design and that knowing basic design techniques can help BI managers deliver data to users in ways that are useful and visually appealing.
That isn't always a skill BI teams possess, though. "If you're building dashboards for a department or enterprise as someone who's running that project, you have to pay attention to design techniques -- something that's alien to IT professionals," said Wayne Eckerson, a BI consultant and director of research for the business applications and architecture media group at TechTarget Inc., the publisher of BI Trends + Strategies.
Story time for visualizing data
The visual design of quantitative data is becoming its own discipline, one that goes well beyond just figuring out how to display graphs and charts in dashboards, Eckerson added. The key principle to remember, he said, is that data tells stories and visualization should emphasize the key plotlines. He advised BI teams to make every pixel count: "Don't just create gratuitous displays and use decorative types of things because you can." Instead, use data visualization capabilities to highlight what business users need to look at in order to do their jobs and make better decisions.
To accomplish that, BI teams need to take into account a variety of specific visualization design considerations. Color, hue, shape and positioning can make important data elements in a chart or other graphic pop, helping users focus on those elements, Eckerson said.
In addition, because the human eye can consume more information graphically than in numerical form, visualizations should convey the required information in a smaller amount of screen real estate than conventional numbers-based reports would need, Eckerson said. Users will get frustrated, he warned, if they have to keep clicking to get the information they need in a visualized application.
The question of sparsity versus density is another key design issue for visualizing data, Eckerson said. "You need to balance those two to tell the story of the data in an optimal way," he said, adding that a BI team needs to know how much data different users can handle at one time. Casual BI users might welcome sparsely populated screens, while more advanced data and business analysts likely will want to see larger amounts of data packed closely together. In both cases, though, Eckerson recommends that dashboards and standalone visualizations should be designed so they don't look cluttered.
About the author:
Christine Parizo is a freelance writer who specializes in covering business and technology issues. Parizo writes for a variety of publications, including several TechTarget websites; she also works as a copy editor for Copyediting, a newsletter for editing professionals. Email her at email@example.com.
- Data Visualization in F1 Racing –TIBCO
- Visual Data Mining and Machine Learning –SAS
- Good Enough to Great: A Quick Guide for Better Data Visualizations –Tableau Software
- Data Visualization Best Practices Guide –Yellowfin