Here is a readable article from Microsoft on what recent OSIsoft Users Conference presenter Bertrand Lasternas at Carnegie Mellon University and other researchers and grad students are doing to improve energy management in their buildings.
"...where things get really interesting is with the potential to do predictive modeling. This is an area where cloud and machine learning technologies have truly been a game changer. Because Carnegie Mellon is collecting and storing real-time and historical data on campus buildings using the PI System, they finally have the ability to do predictive analysis using Azure Machine Learning in exciting ways. "
You can watch Lasternas' presentation and demo at 22:50 in the 2014 EMEA Users Conference session "Big Data and the PI System, Part 1."