Data Visualization Part I: How to Find the Right Tool for Your Business
|Credit: cooldesign via FreeDigitalPhotos.net.|
At its core, data visualization is based on a strikingly simple proposition -- it’s easier to see connections, relationships, and trends when they’re graphically displayed than it is to stare at a bunch of numbers. That realization has created such a demand for data visualization products that it has sparked an “arms race” among the makers of the technology, said one industry insider.
Jim Bartoo, CEO of The Hive Group, which makes data visualization products used in manufacturing, told Tech Trends Journal that “the arms race analogy is entertaining enough, but unfortunately, the arms race actually leads to a lot of confusion and wasted time and resources among the end user organizations, who are supposed to be benefiting from visualization technology.”
Scott Benzie, vice president of marketing and sales for Toronto-based Dundas, advises buyers that abundant features in data visualization products do not equal better solutions. “Features help implement the solution, but they’re not the solution,” he said. “What ends up happening is your vendor tailors the plan towards their own product, and not your company’s problem. This ends up working out quite well for them, and not so well for you.”
James Foster, a marketing manager at SAS responsible for visualization and reporting products, told Tech Trends Journal that data visualization is like any technology implementation in that it “needs to be linked to the business benefit it will bring or the problem it will solve.” Once this wider context is understood, he said, “There are a series of practical questions an organization can ask.”
Foster recommends manufacturing companies be able to address certain key areas before purchasing any data visualization technology. For starters, who are “the information consumers and users” -- the people who use the technology? Understand their technical skills and capabilities, he said. Overshooting will result in an expensive white elephant that benefits no one. “If you limit the process to the vendor and your IT team, you will get a very good visualization that works well for the IT team, but it might not be useful to the actual consumers of the data,” Benzie noted. “This results in low adoption and a lot of wasted resources.”
Foster also recommends knowing how often the information needs to be accessed and from where. Will it need to be used on a mobile device on the factory floor? Know the answer to this before you buy anything to make sure you have something that fits your needs, he said.
What data is being visualized? Is it structured or unstructured? How much data manipulation is required before being able to report on it? Foster said the answers to these questions will point the company to what it actually needs.
It’s also important to know the ratio of reporting types -- including basic reporting, scorecards, dashboards, operational alerting, data discovery, forecasting and predictive analysis, he said, adding “Is there a need or benefit in visualization of the data geographically?”
Have these questions answered before sitting down with vendors; don’t let them answer these questions for you, he advises.
That being said, Bartoo pointed out that vendors can help companies think through what’s possible. “A lot of companies don’t really know what they want out of a visualization solution, or what types of functionality they’re going to require,” he said. Engaging with a vendor early on means “their requirements can be, in part, driven by what’s possible.”
Integration is a key issue. “One of our customers referred to another vendor as really just offering collections of pretty pictures, referring to their total lack of consideration for integration into existing environments and infrastructure,” Bartoo said.
Benzie suggests that buyers allow the vendor to be the expert in visualizing the solution, “but don’t lose sight of your problem throughout the process. It may sound obvious, but it’s been proven time and time again that companies get caught up in the allure of features, and not whether or not that vendor is equipped and skilled enough to handle the company’s problem.”
Foster says that when SAS sits down with companies to discuss data visualization, they find companies are also interested in such practical questions as how easy is the product to use for non-technical users, whether it can access multiple data sources from across the organization and externally, whether it can scale to very large data volumes -- “especially at a transactional level over a long period of time” -- and whether an unstructured data, such as warranty claim forms or product descriptions, can be incorporated into the visualization environment.
It’s all about problem-solving, Benzie said. “When approaching any vendor, don’t ask them for a bubble chart,” he said. “Ask them about their process for discovering the right visualization to solve the issues at hand.”