Data visualizations might be a great way to deliver analytical insights to business users, but that doesn't mean you load up the visual dashboard and get out of the way.
Sophisticated data visualization tools like SAS Visual Analytics need professional drivers, recommended Kimberly Holmes, head of strategy analytics at XL Group, a global insurance and reinsurance company headquartered in Dublin, Ireland. "At XL, data visualization is going to be a team effort -- we'll have the analysts in the driver's seat and the business owner in the passenger seat." (See: Visualizing What Questions to Ask of the Data for more information on XL's plans for Visual Analytics.)
Holmes's strategic analytics team already works closely with the users. "In fact," she said, "I always warn businesses how much time they'll be spending with us and they're still always surprised." But Holmes said she anticipates Visual Analytics helping to change the nature of that relationship.
For one, it'll help overcome the common struggle of advanced analytics teams in any industry: "finding analysts and data miners and modelers who have domain expertise." Visual Analytics will allow the analyst to sit with the business owner and steer while the business owner calls the directions -- "I want to look at this and I want to look at that," Holmes said.
And with this interplay, the gap between the analysts, who don't have insurance domain expertise, and the business owners, who aren't analytical experts, starts to close, she added. "It'll level the playing field and improve the communication between the analysts and business owners -- which is key to the success of analytics in general."
The analytics team that stays in its ivory tower, developing tools and insights apart from the business, is placing itself in a precarious spot. "We have a) execution risk and b) implementation risk -- the business users may not want to embrace what we've done if they don't really understand it. By working together, using Visual Analytics, we'll have a real team effort and each side will get a deep appreciation for drivers of success for the other side."
Holmes said she'll take a multipronged approach in working with the business managers, first inviting them to a presentation on how her team can help their team and giving them a data visualization demo. They'll be blown away, she said she's convinced, just as she was when seeing Visual Analytics in action for the first time while visiting SAS headquarters in Cary, N.C.
"I wish I'd had a tape recorder in my head so I could have recorded every idea generated during my demo. Each time we looked at the data, I thought, 'Wow, I never thought about it that way, and we should be asking these extra three questions' " -- and this wasn't even with XL's own data. The demo was that effective, Holmes said.
After all, "It's all about asking the right questions, because if you don't ask the right questions you don't ever get the answer and the insight."
XL's strategic analytics team will build up its internal data visualization capabilities with a 50,000 dataset recently used in a univariate analysis with 270 variables. Once comfortable, it'll load up a recently purchased larger-scale dataset for exploratory analysis with the business. That larger-scale dataset has millions of rows, Holmes said.
"Maybe we'll do some teasers, pointing out interesting things we found, but then we'll let the business leaders ask the questions and drive the discussion. They obviously understand the business better than we do as analysts."
The data visualizations are sure to open a few eyes. "There are a lot of people who understand theoretically the possibilities that the analytics presents. But what I'm hoping is that when we sit down with the businesses, we'll make the information extremely tangible for them and garner lots of enthusiasm."