Hi, I am Tim, a doctoral student at NYU Shanghai and NYU Wagner, researcher at the Shanghai Key Laboratory Of Urban Design and Urban Science. My advisors are Professor Guan Chenghe and Professor Guo Zhan. I received my Master's degree in Data Science from Columbia Data Science Institute and Bachelor's degree in Data Science and Finance from NYU Shanghai. Previously, I worked at the Center for Spatial Research at Columbia University, where I lead the data science work on the Mapping Historical New York project. My research interest is to create and analyze high resolution urban digital twins using machine learning and GIS, so as to design more sustainable, resilient, and inclusive cities. Below are a collection of my previous projects organized by theme.

• View My C.V.

Sustainable and Green City

Capture Pedestrian Flows with Traffic Cameras

Detect pedestrians in 18 million photos of street intersections and identify patterns with time series clustering.

Evaluate Reliability of Green View Index in NYC

Identify challenges in measuring Green View Index by comparing it with tree census, lidar, and field observations.

Detect Forest Changes with Remote Sensing

Develop robust high-resolution forest change detection based on Dynamic World and remote sensing data on GEE.

Urban Greenery and Excessive Heat Waves

Study the relationship between extreme air temperature and urban greenery in Changjiang delta region with GWR.

Pandemic Resilience

Built Environment Factors and the Pandemic

Examine the impact of street network and density on the pandemic through regression analysis of a global dataset.

Accessibility and Equity of COVID-19 Tests

Analyze the evolving distribution of NYC test sites and their spatial accessibility with 3SFCA.

Explain Uneven Recovery of Subway Ridership

Build high spatial-temporal resolution data to model ridership at different stations with XGBoost.

Optimize Subway Network Passenger Flows

Model subway network with turnstile data and propose optimization that balances traffic time and infection risk.

Data-driven Urban History

Historical Urban Model through Map Extraction

Scalable pipelines to extract map features from scanned historical maps, leveraging CV and NLP techniques.

Building-level Land Use Evolution Model

Buildings and streets of Manhattan and Brooklyn in 1859, with detailed land use and building use information.

Longitudinal Tax Lot Matching for Geocoding

Match historical tax lots from 1940 to modern MapPLUTO dataset for the reconstruction of historical address system.

Manhattan Immigrant Heatmap

Heatmaps of major immigrant community in 1910, overlaid with occupation, language, and other demographic information.

Open-source Tools

Urban Open Data Recommendation Engine

An AI-powered search and recommendation platform that helps you discover and connect open urban datasets.

Place Type Extractor for Fuzzy Matching

Automatic type extraction from facility names with a language-agnostic, bottom-up approach.