The creation of prosumer communities leveraging the smart grid domain to improve energy production and consumption patterns has been proposed in the literature. We propose to use a novel Internet of Things (IoT) paradigm, namely, the Social IoT (SIoT), to enable appliance-level metering (submetering) and to create prosumer communities dynamically. We use the term Social Internet of Energy (SIoE) to refer to the integration between prosumers and devices via social relationships. This article explores the diverse technologies that will enable the SIoE. On the one hand, we explore the different aspects and value propositions of the SIoT, applying this novel paradigm to the creation of prosumer communities. On the other hand, we explore demand side management, its synergy with SIoT and the properties and challenges that may arise. Concretely, in this article, we aim at grouping prosumers by their energy load shape compatibility and then reschedule intragroup load shapes. We propose a clustering model and we develop two clustering heuristics to distribute prosumers among prosumer community group and one peak-to-average ratio (PAR) prediction heuristic in order to reduce the PAR and the amplitude of energy peaks per cluster. Then, on the basis of a real dataset, we perform simulations and analyze and discuss the results, paving the way for further research.