Improving the energy performance of educational buildings is important for the promotion of an energy efficient culture among future generations. The examination of energy usage with maintained comfort levels is especially relevant for buildings like schools and colleges, where the occupancy has high variability within small time intervals and often periods of low to zero occupancy. This variation in occupancy is currently not taken into consideration on campus with the majority of classrooms run on a set heating schedule. With public sector bodies expected to reduce energy consumption by 33% by 2020 under Ireland’s National Energy Efficiency Action Plan, there is an onus on universities in Ireland to reduce their energy consumption. This work aims to evaluate the benefit of incorporating timetable and occupancy information. The deployment of an adequate number of sensors and meters in retrofit applications can be cost prohibitive with installation costs alone accounting for 70% of the costs. Data gathered from the deployment of cheap off the shelf sensors in combination with occupancy information obtained from the schools timetable information resulted in a reduction in heating use of 25%. The incorporation of the j48 tree as a machine-learning technique helped create a decision network with 90% accuracy that could inform a new heating system that is responsive to current need.