Data, models,
and fieldwork.

I studied where San Diego County has gaps in air-quality monitoring and how coverage compares with population density.

With guidance from University of Pennsylvania Professor Davi Maximo, I built a Python pipeline to map PurpleAir sensors to census tracts and adjust readings for humidity and county-wide pollution trends. I then used Ollivier–Ricci curvature to find bottlenecks in a network based on sensor locations and correlations.

Analysis: 395 sensors and 54,354 PM2.5 and humidity records mapped across 199 census tracts, producing 40,896 daily tract-level observations.

The 18-page paper is accepted for publication in Innovation in Science and Technology.

Read the paper ↗

I worked on sample preparation, calibration curves, and analysis of seven major ions.

Chloride and sulfate in the tested samples exceeded the limits in China's GB 5749–2022 drinking-water standard.

I wrote a 27-page report covering our methods, results, and comparison with drinking-water standards.

I join weekly research discussions and analyze global and San Diego temperature records in Python and Jupyter. My work uses moving averages, linear regression, maps, and basic climate models.

I tested whether models trained on older cars could predict fuel efficiency for newer ones.

At UChicago, I coordinated a four-person team studying nearly 400 cars with k-nearest neighbors and multiple linear regression. After the program, I added newer vehicle data and EPA trends through 2024 on my own.

The project includes six notebooks, a repeatable data pipeline, a report, and a website.

Explore the analysis →

I studied Chinese American history in seminars and through fieldwork in San Francisco Chinatown.

I interviewed eight local business owners and residents, photographed signs, family businesses, and cultural spaces, and wrote about how traditions change across generations and migration.

My photos, interviews, and cultural analysis were published in Chuan — The Red Thread.