The Data Lake Versus the Data Warehouse

The data as it moves along is becoming in a format that's much more easier for end users to consume. Some customers are introducing machine learning algorithms to say if the count seems out of the ordinary, if it's an outlier. And so I've seen some pretty sophisticated ways of trying to trigger at least a email alert that somebody should go look at this because it doesn't seem right. The only way you can do it is spend a lot of time upfront asking those questions and then seeing what's going to work.