Data visualization is one of the hottest capabilities sweeping over the big data market. Visualization platforms sit on top of analytics and present information about organizational data using a barrage of colors, heat maps and word clouds. In the midst of all that dazzling information, however, a key question remains: what’s going on under the surface?
As companies flock towards visualization, it’s important to remember that a great visualization alone does not translate into great insight.
Organizations often assume that analytics and visualizations are synonymous. In reality, they’re vastly different. Think of it this way: data and analytics are like the foundation and structure of a house. Visualization is the exterior aesthetics – be it great trim, a coat of paint or beautiful landscaping. The exterior of a home can look amazing, but without a solid structure, the home will never pass inspection. Similarly, if not grounded in proper analytics, data visualization crumbles once you dig below the surface.
Too often today, people see a visualization and assume any number of things – that the underlying analytics are sound, that supporting data is valid, and that the right questions are being asked to begin with. But do we really know for certain? Think back to the last visually compelling presentation of data you saw. Do you know for sure that all the data sources needed to make the analysis relevant were included? Do you know for sure the right data from the right time periods had been pulled for the analysis? Most importantly, were you confident enough to make a key business decision based on what you’d just seen?
Take the general example of word clouds, a popular form of visualization. Dazzling to the eye, they’re more often than not an empty visualization, largely because their presenters rarely explain the supporting data or analysis, and their consumers are rarely able translate them into an actionable business insight. So the presenter displays the word cloud, the reviewer nods in approval, but neither party has enriched their understanding of the business. Perhaps not the best use of precious corporate resources.
To put it in real world context, as we do with all of our customers, Dell Software’s information management team recently worked closely with the team Kennesaw State University, in this case on an analysis of the company’s HR data.
The team at Kennesaw had been using visualizations to show certain data trends, but when we enabled them with greater self-service data access and analytic exploration by way of Dell Software’s Toad Business Intelligence Suite, they realized certain key data sources were missing from their analysis. As a result, they had been overpaying their contractors for the past two years. Those visualized reports were in fact misleading the organization because the data analysis behind them lacked the robust analytics required to paint a complete view of the operations.
Today, Kennesaw has the ability to connect to and analyze all data, as well as the ability to present that data in form of compelling visualizations. As Kennesaw’s Brad Smith, director of payroll services and human resource information systems noted in a recent Dell case study, “We have multiple data sources, and we’re finally able to tie all of them together with this solution.” With those capabilities in place, Kennesaw can confidently accomplish key HR objectives such as paying competitive salaries, reducing employee turnover, and ensuring the profitability of summer sessions.
The takeaway for customers is simple: The quality of a visualization is secondary to the quality of the analysis behind it. If you’re presenting a visualization, become intimately familiar with the supporting data analysis. On the other hand, if you’re consuming a visualization, ask questions. You can still say, “Wow, this looks great.” Just be sure to follow that up with, “What metrics are represented here? What are the time frames for these analysis? What is your confidence factor in the accuracy of these visualizations?”
Remember, just like the house, where a glossy coat of paint on the top can cover a massive structural flaw, so too can visualizations that aren’t rooted in sound data analytics mask structural flaws in your business processes. That’s why Dell Software has prioritized building a complete portfolio of information management solutions, one that includes BI, data discovery and visualization tools such as those in the Toad Business Intelligence Suite, but also the critical predictive analytics tools recently brought into the company through the acquisition of StatSoft. Visualization has a place, but its analytics – supported by the ability to connect to all data – that drives your business.