Scope
The policy project combined qualitative interviews with broader transportation data. That meant the work needed to stay readable for policy authors while still being structured enough to support evidence-based claims.
What I Did
I developed a Python NLP workflow using KoNLPy, word clouds, and regex-based preprocessing to extract themes from FGI data with pregnant women and caregivers. I also worked with nationwide datasets related to rural transport routes, urban mobility models, and public taxi subsidy programs.
Why It Was Useful
This project mattered because it forced multiple forms of evidence into one usable narrative: interview findings, comparative case studies, and structured datasets. My role was to help make those pieces easier to clean, analyze, and translate into the final report.