In my PhD I am focusing on how to support householders with solar PV increasing their self-consumption (consumption of the energy coming from their panels). To do that I am investigating energy forecast. Although there are existing works about that, most of these solutions use MathLab or other big engines that are not really convenient in the real time and distributed context. I am looking for tools, techniques, technologies that could provide a prediction accurate enough without a lot of effort from the system.