Bibliographic Review of Data-Driven Methods for Building Energy Optimisation

Abstract

This study presents a systematic bibliographic review of the application of Big Data and machine learning (ML) methods to improve energy efficiency in architectural design. The review covers peer-reviewed publications from 2010 to 2025, examining how ML algorithms such as Random Forest, Gradient Boosting, and neural networks have been used to optimise design parameters including orientation, glazing ratio, and compactness. A systematic search and selection protocol was applied to identify, classify, and critically analyse over 70 relevant studies. The findings reveal consistent evidence that data-driven models outperform traditional simulation-based methods in predicting heating and cooling loads while highlighting current gaps related to data quality, model interpretability, and real-world validation. The study contributes to the understanding of how ML-driven approaches can guide sustainable architectural design and future research directions in the built environment. Additionally, illustrative experiments were performed using simulated datasets to validate and exemplify key findings identified in the reviewed studies. © 2025 by the authors.