Spotify Playlist Analysis
Collected and wrangled data from Spotify Web API regarding its genres, playlists, tracks and audio features
Built regression models to predict playlist popularity
Compared model performance and constructed stacking model of the best performers
Data at Large - A Visualization Project
Acquired open-source data sets revolving cybersecurity, including internet usage, company data breaches and identity thefts
Created a interaction visualization website aimed to raise awareness on cyber-crimes
2017 NBA Hackathon Entry - New Boxscore Statistic
Analyzed five years of NBA game results to gain insights of underestimated aspects of player performance
Developed three innovative game statistics that are well justified and influential to the game
Cancer Tumor Classification
Researched deep learning models for image classification
Transfer learning of Inception-ResNet, a Convolutional Neural Network
Undergraduate Capstone Project: Intelligent Usage of Market Smart Data
Helped a semiconductor company improve their demand forecasting model
Implemented state-of-the-art models on given dataset and benchmarked their performance
Built an ensemble model that improved accuracy by 24%
Developed a Python application that automates demand forecasting, visualization and reporting
See project report / See project poster
Urban Freight Transport Data Analysis and Visualization
Studied urban freight movements by carrying out exploratory analyses on datasets containing traffic flow information
Identified patterns in Singapore vehicle movements through data visualization
Network Analysis of Global Visa Requirement
Scraped the web to collect data on visa requirements of more than 150 countries
Conducted network analysis of the data to understand its pattern and its implications in global relationships