Data-Driven Human Perception Modeling for Sustainable Architectural Design

You are cordially invited to attend the lecture by Kim Jaeha:

Date: 7 Oct 2026
Time: 10:00 AM – 11:00 AM
Venue: SDE3 Level 4, LT424

Data-Driven Human Perception Modeling for Sustainable Architectural Design

Dense urban housing must balance daylight and outdoor views with protection from unwanted visual exposure, yet conventional performance workflows rely largely on geometric or compliance-based indicators. This research develops a human-centered computational framework to measure, model, and scale residential visual comfort across daylight, window view, and visual privacy. Human-perception data are collected through physical, virtual reality, and image-based studies, then translated into predictive models using CAD-based ray casting, semantic segmentation, depth estimation, and machine learning.

A Visual Privacy Index converts geometric exposure into perception-based performance. Validated in New York City and scaled to 35 buildings and 10,420 rooms, the framework reveals how daylight, views, and privacy vary with floor level, window configuration, and urban morphology, supporting design, analysis, standards, and data-driven housing decisions.

Jaeha Kim is an architect, computational researcher, and Ph.D. candidate in Systems Engineering at Cornell University. Her work lies at the intersection of architectural design, building technology, human perception, and artificial intelligence, with a focus on developing human-centered computational methods for evaluating and improving the built environment.

Her doctoral research investigates daylight, window view, and visual privacy in high-density urban housing. By integrating empirical occupant data with environmental simulation, computational geometry, computer vision, and machine learning, she develops new performance metrics and predictive models that translate human experience into actionable design information. She is also the co-founder and lead developer of Viewscore.io, an AI-enabled platform for predicting occupant satisfaction and supporting design decision-making.

Trained in both architectural design and systems engineering, Jaeha previously practiced as a senior designer at BAU Architects in Seoul before pursuing graduate research at Cornell. Her broader research agenda explores how emerging computational technologies can support healthier, more sustainable, and more responsive architecture, particularly in the context of increasingly dense urban environments.