With the leap of computational and storage powers and the evolution of sensing and communication technologies,  massive amount of data is being generated, transmitted, and retained in near real-time. All of this contribute to the large-scale growth of exploitable, and often open data, pose a significant opportunity, and trigger disruptive changes across numerous fields including energy management in smart buildings and smart cities. However, extracting knowledge from this data faces several challenges. The most obvious challenge is to process the large-scale voluminous data by relying distributed and parallel computing frameworks. In addition, collected big data are most often multimodal and comes from multiple heteregenous data sources. Thus, it requires heterogenous data integration solutions. Furthermore, on-line analytics required in order to process high-speed streaming data on the fly.

The purpose of this research project is to investigate the above-mentioned challenges while examining the case study of smart building's big energy data processing and analytics.

Mots clés : big data analysis, domain adaptation, machine learning, building energy modeling, smart building.