​Segment for Solar Facades (S4S) is a deep learning–powered tool that analyzes building façade images to identify suitable areas for photovoltaic installation, integrating solar irradiance data to estimate energy yields with architectural detail.​

This review analyzes trends and applications of AI methods—including Machine Learning, Deep Learning, and Computer Vision—in facade research, highlighting current uses, limitations, and future potential in architectural, environmental, and structural contexts.

This study investigates how building-integrated photovoltaic (BIPV) façades affect outdoor thermal comfort across four urban block typologies in Zurich, revealing that urban form and climate change have a far greater impact than BIPV installation.

This work presents a method that combines open data sources and deep learning to extract detailed architectural façade features for more accurate estimation of BIPV energy potential.

This study introduces a deep learning–based method using GANs and street-level imagery to accurately estimate BIPV façade surface utilization, showing significant discrepancies with existing tools and emphasizing its potential to inform better solar energy planning and policy.

This study uses a parametric simulation approach to assess how varying levels of BIPV façade deployment influence pedestrian-level outdoor thermal comfort across different urban settings, highlighting that factors like ground material and street width have a greater impact than BIPV itself.

This study explores how different urban street network patterns influence the BIPV generation capacity of buildings, showing that irregular urban layouts enhance façade PV potential while roof PV potential remains largely unaffected by morphology.

This study enhances urban building energy models (UBEM) by integrating open data on occupancy and residential layouts, demonstrating that data-rich models significantly improve the accuracy of energy use and thermal comfort..

This thesis proposes a machine learning–based, data-driven method to predict BIPV generation potential on façades and roofs in diverse urban contexts, demonstrating high accuracy and highlighting the influence of urban morphology on solar performance.

This study uses data-driven decision tree models to analyze how varying occupant thermal preferences impact comfort and energy performance in naturally ventilated classrooms, revealing key tradeoffs between occupant well-being and resource consumption in educational buildings.

This study develops artificial neural network models to accurately predict heating demand, indoor overheating, and CO₂ levels in naturally ventilated classrooms, offering a fast, data-driven alternative to traditional simulations for evaluating classroom design and retrofit scenarios.