I seek to leverage the power of data science, visualization, and web development
to lead better urban planning, public policy making, and business growth.

Master of Urban Spatial Analytics'20 @ the University of Pennsylvania
Currently working @ Meituan as a data analyst.

Predictive Modeling and Analysis



Transit Ridership Prediction and Scenario Building in Austin

Skills used: R, Random Forest, XGBoost, Scenario Testing

Train Occupancy Prediction in Belgium

Skills used: R, Multinomial Logistic Regression, App Wireframe, Profit Optimization

Eurosat Image Classification with Deep Learning

Skill used: Python, Keras, CNN, VGG16

Food Desert Prediction in New York

Skills used: Python, Web Scraping, Twitter API, Sentiment Analysis, Street Network Analysis (Pandana), Machine Learning (Scikit-learn), Data Visualization (Altair, HvPlot, Seaborn, Folium, Panel)

Domestic Batteries Prediction in Chicago

Skills used: R, Poisson Regression Model

Home Prices Prediction in San Francisco

Skills used: R, OLS Modeling, Spatial Lag

Architectural Form Finding via Machine Learning

Skill used: Python, Tensorflow

Rent Price Analysis Scraping from Craigslist

Skill used: Python, beautifulsoup

NDVI Analysis in Philadelphia

Skill used: Python, rasterio, rasterstats




Web Application & Visualization

Covid-19 Cases Visualization

Skills used: Javascript, Leaflet, HTML, CSS

Austin Transit Ridership Prediction and Visualization

Skills used: Javascript, Leaflet, Billboard, HTML, CSS

Visualize Parking Violations in Philadelphia

Skill used: Python, datashader




Writings

Prediction of Median House Values in Philadelphia Block Groups

Predict Philadelphia house values with spatial lag, spatial error and geographically weighted regression (GWR).

Argentina Slums Identification with Deep Learning

Skill used: Python, Machine Learning (Scikit-learn), K-Means, SVM, Deep Learning (Keras), Convolutional Neural Network (CNN), ResNet

Forests Patches Preservation
Using High Carbon Stock Approach

Find the viable forest patches that are of the most conservation importance through patch analyses. Skill used: Google Earth Engine

Recidivism Memo

Evaluate the fairness of algorithms in predicting recidivism.




Architecture and Landscape Design

About Me