Continuous data applications are generating a lot of buzz for their ability to move and shift along with real-time data flows. But it’s not just what the application can do for your data; it’s what the data can do for your application. The best thing about these new apps is that you don’t have to scrap them and start over from scratch when new data comes in. But that doesn’t mean you can’t use the new data to make smart changes.

Chris D’Agostino, VP of Technology at Capital One Financial Corp., spoke about his company’s use of Spark tools for combining data sets. He told hosts John Walls and George Gilbert (@ggilbert41) of theCUBE, from the SiliconANGLE Media team, “We’ve got a lot of insights that are generated from our historical data,” and explained that that data could still be operated on in the old batch model.

He views new streaming data as highly valuable and said, “We want to marry up the two, take the insights and profiles that we’ve been able to glean out of the historical information and couple that with the inbound information and try to extract the features as quickly as possible.”

The streaming data twist

D’Agostino, said that while the resiliency of the streaming application isn’t lost on them, they still want to utilize the streaming data to “inform the validity of the model and improve the model over time.”

He said that running models in parallel in an A/B test would help them improve them over time.

Watch the full interview below, and be sure to check out more of SiliconANGLE and theCUBE’s coverage of Spark Summit 2016.

Photo by SiliconANGLE