Microsoft has a great customer story on how Carnegie Mellon reduced their building maintenance and energy costs using Azure and Azure Machine Learning, or simple “Azure ML.”

The Azure ML component of the overall solution gave them a way to predict (and prevent) imminent building system failures, by repairing or replacing aging and worn components before costly failures. And it also enabled automated building systems to predict when to adjust thermostats and by how much, leading to energy cost savings.

By looking at the historical building operations data created by their OSISoft PI System, and discovering complex patterns resulting in a predictive model that must first “predict the past” , Carnegie Mellon can apply that predictive model to current, real time building data and “predict the future”, optimizing their building operations and energy costs.

I like this story because it’s a great example of the breadth of possible applications for getting more from your data with analytics. Not all business applications for predictive analytics are just about marketing and sales, as you can see from this example with Carnegie Mellon.

Any business process that creates data and costs you a sizeable pile money is a great target for doing more with that data via analytics, aka “Business Analytics”. You can keep it simple as you get started by focusing in using basic descriptive statistics (e.g. mean, median, standard deviation, histograms, scatter plots, etc) to look more closely at what your data might be trying to tell you. This capability would be the first stage of Business Analytics, typically referred to as “Descriptive Analytics”. After that, you can press on to the second stage of Business Analytics, referred to as “Predictive Analytics”. Utilizing machine learning technologies like Azure ML or Amazon ML, you can discover complex patterns in your historical data that result in predictive models you use to make better, predictive decisions now and in the future.

If you can create and nurture a culture of questions, experiments and data… and if you can teach your people how to accept, understand and use what predictive models are telling them… you can harness the full power of Business Analytics and become a potent analytical competitor in your industry.

You can read more about Business Analytics here in my blog.

You can read the Carnegie Mellon story here on the Microsoft website.


Gene Connolly is an independent consultant with a passion for data and analytics; reading, writing and sharing what he knows with others who have an interest in how they can harness the power of prediction for themselves. You can find him on LinkedIn , on Twitter  or by email at