I think the Anamoly Dection feature is very cool (still wish you would bring back Forecasting, but that's another post). The Training Period needs some functionality to make it more useful. As another user mentioned, being able to be more granular like excluding weekends or weekdays would be helpful. More than that though, for organizations that run their business seasonally, its not helpful to detect an anamoly between November 2013 and a training period of August-October 2013 because what we want to compare against is the same time period the year before.
So the enhancement request is to be able to customize the date range with more options. I also think it'd be helpful if we could add this to dashboard or be able to email this report to others.
