We employ a number of models and methods to determine the most suitable for each site, with the best R squared value, either through multi element linear regression or ARIMA (Auto Regression with Integrated Moving Average). From this it is possible to calculate both a baseline, with normalised consumption prediction based on actual conditions, and projected budget figures using 10 year average conditions. 

The regression analysis is used to identify and quantify project savings with CUSUM (cumulated saving) analysis as well as in the preparation of advance budgets.