Adaptive Façades for Activity-Specific Well-Being:Exploring emotional responses with virtual reality and machine learning

While façades shape interior ambiance through light and shadow, architects lack evidence-based methods to tailor dynamic façades to activity-specific emotional needs. This study tests a novel framework combining Virtual Reality and unsupervised Machine Learning to identify how users configure façade features (speed, coverage, spread, position) for distinct activities (working, relaxing, socializing). Seventy-five participants customized façades in an interactive environment and rated their emotional states. Analyses revealed both activity-driven preferences (slow-moving, low-coverage façades for work/relaxation vs. dynamic, high-coverage patterns for socializing) and participants’ cluster specific trend. Although these findings need validation across demographics, they demonstrate that Virtual Reality experimentation and Machine Learning clustering can decode activity-specific preferences, offering architects a data-driven starting point for adaptive façade design.