Abstract: Energy efficiency is still a hot topic today. Coming roughly the 25% of the energy consumption in EU from the residential sector, very few cheap and simple tools to promote energy efficiency in home users have been developed. The purpose of this paper is to present Bernard, a concept proof designed for filling this gap. This aims that householders become aware of their energy habits and have useful information that help them to redirect their consumption pattern. To achieve these goals, Bernard offers, through a mobile application, the home energy consumption monitoring in real time, the energy price forecast for the next hour and the appliances which are switched on, among others. Furthermore, it is important to highlight that the system has been designed with the premises of being cheap, non-intrusive, reliable and easily scalable, in order that utilities can gradually deploy and provide it to their customers, gaining at the same time valuable information for decision making and improving its corporate social image. Therefore, the adopted solution is based on a real time streaming data architecture suitable for handling huge volumes of data and applying predictive techniques on a cloud-computing environment. The paper provides a detailed description of the system and experimental results evaluating the performance of the predictive modules built. As case study, REFIT and REDD datasets were used.
