As buildings become increasingly automated and energy efficient, the relative impact of occupants on the overall building carbon footprint is expected to increase. Research shows that by changing occupant behaviour energy savings between 5 and 15 % could be achieved. The definition of occupant behaviour is however highly dependent on the context, and how to successfully develop behaviour change interventions is an open problem. This thesis specifically focuses on the ineffectiveness of Smart Meters based on visual feedback, analysing the reasons behind their faults and exploring alternative feedback methods based on sonic interactions. This work is an investigation of sonic interaction design methods for energy awareness, based on quantitative data analysis of real-world datasets of historical disaggregated energy consumption.