An integral part of contemporary initiatives striving to promote and create self-sustainable communities has been the integration and efficient management of renewable energy sources, energy storage solutions, and sustainable heating and cooling assets. This paper presents a unique combination of diverse technologies, ranging from established data management solutions like SQL and NoSQL databases to custom semantic solutions and device-specific control adapters powered by the OpenMUC gateway. Additionally, it incorporates analytical machine learning-based forecasting solutions paired with optimization algorithms, working together to enable energy self-sufficiency. The platform constituted by these solutions is subsequently utilized to provide predictive and real-time control of energy assets in various facilities in line with the selected operation strategy. Its application, particularly in terms of effective energy storage utilization and timing of asset activation scheduling, ultimately results in improvements in renewable energy integration and overall increase of energy efficiency in the considered buildings. As demonstrated in real-world use cases tested within prosumer-based energy communities in geographical islands, the application of the platform resulted in tangible modifications to the primary statistical characteristics of electrical energy consumption clearly signified, among others, by a reduction in mean consumption of between 40W and 90W. Through scrutinizing the achieved results, it can be concluded that the platform displays the capabilities to outperform stock control algorithms provided by the inverter vendors. Expanding upon the application limited to the electric assets, use cases in the thermal domain where optimization outputs are utilized for heat pump scheduling were also discussed and presented.