The approach we took was to investigate similarities in behaviour for energy usage across similar type of sites whilst accounting for seasonal change. By comparing institutes to their own history has allowed us to get a good idea of when their behaviour is unusual and should be investigated for potential savings.
The outputs from this was a dashboard document which shows just a small amount of the information that can be gleamed from the data available and by using exactly the same data and approach a notification system can be implemented to get quick notification for issues.
The solution was built using an Azure SQL Database for data storage and transformation. Tableau was then used to visualise this information. From there a dashboard was created to show the real insight from the data.