Hey, I'm Chaeyeon Han

🖥️ Data Centers, Housing Prices, and Migration

Estimating how the arrival of large-scale data centers relates to local migration rates and housing prices, using consumer address histories and housing transaction data.

Ongoing

Image source: Climate Camp Ireland

🧳 From Evacuation to Relocation: Hurricane Mobility Sequences

Dissertation research using individual address-to-address migration histories to trace how Hurricanes Helene and Milton turned evacuations into longer-term relocations, testing whether social ties and socioeconomic homophily explain where climate migrants settle.

Paper under review · ACSP 2026 presentation (Pittsburgh)

🌀 Can We Learn from Prior Hurricanes?

Tests whether evacuation destination choice models transfer across three hurricane events, using mobile GPS-derived (Advan) foot traffic data and explainable AI to identify what drives destination choice, with a focus on social ties.

Paper under review · AAG 2026 and Multiplied Displacement 2026 presentations

🌊 Where Do Evacuees Go? Destination Choices during Hurricane Milton

Used mobile GPS-derived (Advan) visitation data and quantile regression to trace where Florida residents evacuated in Georgia during Hurricane Milton. Social ties drew evacuees to destinations, while distance from urban centers and proximity to airports discouraged them.

Publication (International Journal of Disaster Risk Reduction, 2026)

🏙️ Climate Migration Indicators: Cincinnati Case Study

Bibliometric review of climate migration indicators and forecast models, followed by projections of the number and profile of climate migrants moving to Cincinnati using public data (IRS SOI, ACS, PUMS, FEMA, NOAA). In collaboration with the City of Cincinnati and ICLEI.

Report with Karen Chapple, School of Cities, University of Toronto (2026)

🌦️ Does Weather Matter in Street View Perception?

Examined weather-induced measurement bias in street view image perception scores across Los Angeles, Seattle and Portland, with guidance for valid perception analyses.

Publication (Computers, Environment and Urban Systems, 2026)

🎧 Pedestrian Sensing with Audio and Video

NSF-funded project (award 2203408) testing how far low-cost microphones can be pushed to detect and count pedestrians.

• ASPED audio pedestrian dataset (ICASSP 2024)
• Audio pedestrian detection under vehicle noise (DCASE 2025)
• Predicting future pedestrian flows from audio and video-based detection data (Urban Informatics, 2024)

Publication Project Website USDOT Report

💡 Teleworking and Energy Use

NSF-funded project (award 2243100) linking telework, e-commerce, time use and energy.

• Residential energy: multilevel modeling and time use comparisons of teleworking frequency and home energy use in the U.S. Publication (ERIS, 2025)
• Vehicle use: path analysis of the 2022 National Household Travel Survey. Frequent teleworkers live farther from work, yet households drive about 10% less per teleworker, roughly a 0.4% cut in national transportation emissions. Publication (ERIS, 2026)

📄 Plan Evaluation using ChatGPT

Assessed ChatGPT as a plan evaluation tool, specifically for scoring plans with a flood resilience scorecard.

Publication (Journal of Environmental Planning and Management, 2026)

🍝 Aesthetic Places, Streetscapes, and Place Satisfaction

Assessed indoor aesthetics and streetscape safety with computer vision models, combining SafeGraph POI and sales data, Yelp text and images, and Google Street View to test whether streetscapes still matter in car-dependent neighborhoods.

Publication

🍁 Can Thanksgiving Destinations Predict Climate Migration Patterns?

Compared post-hurricane relocation destinations with Thanksgiving holiday travel patterns in the U.S., using SafeGraph weekly foot traffic data and Spearman rank correlation.

Publication

🏡 Mental Health and Built Environment

Mediation analysis of how walkability relates to mental health outcomes, using SafeGraph visits to mental health facilities in Atlanta.

Publication

😷 A CyberGIS-enabled Pandemic Warning System Using Social Media Data

Participated in creating CyberGIS-Viz, a geovisualization tool for showing spatiotemporal dynamics in infectious disease datasets, alongside COVID-19 risk prediction (ARIMA, LSTM and Prophet) using Twitter-derived mobility.

Project Website Publication