Pat Pat Cat - Mobile Game, 2022
Cat themed whack-a-mole style arcade game. Play through 6 maps and 30+ levels. Collect all 24 unique cats. This project was a valuable experience to design and develop a product end-to-end.
I worked in collaboartion with Cindy Chang who worked on the art and audio. I worked mainly on the programming, economy design, and data analysis. We collaborated on the game design.
Pat Pat Cat Play Store Link
Pat Pat Cat website
Cribbage Assistant, 2017
An app where cribbage players can score and optimize their hands. It is also useful for beginners and advanced players to brush up on their skills.
Cribbage Assistant App LinkProductivity Beats, 2016
A chrome extension that let users block time consuming sites and play their own music files.
Chrome add-onElectric Vehicle Traffic Intensity Estimation, 2015
A statistical consulting project examining the expansion of the current infrastructure of electic vehicles at UBC. The goal is to determine the number of additional charging stations to construct in order to meet the demands of increased electic vehicle usage in the future. We use queuing models in the analysis, and a visualization app for illustration.
Repository | Report | Visualization App | PresentationM.Sc. Project, 2015
This project involves the development of tree-based supervised methods to discern different waveform types from radio signals transmitted from wireless communication systems. The tree-based methods include classification tree, random forest, and feature-based tree.
Repository | M.Sc. Report | Paper
Random Forest Visualization Applications, 2014
I developed two visualization applications as part of the final project of CPSC 547: Information Visualization. As random forest belongs to a class of black-box models for machine learning, the purpose of the visualizations is to provide intuition and descriptive information for the set of trees generated from a random forest classifier.
Indented Aggregate Tree
The first application is built with python, d3, and javascript. It displays aggregated feature variables at node positions based on the decision trees generated by a random forest classifier.
Full article | Visualization | Repository
Ensemble Quantification and Filtering
This application fits a random forest classifier on a fixed data set, and filters the ensemble of decision trees of the classifier by quantifying the trees by hamann similarity measure. The application is implemented in RShiny.
Social Network Analysis, 2014
This social network analysis involves the community detection of a network of friends on facebook. The image on the right shows that the walk-trap community finding algorithm discerns most of the relationships correctly.
Stochastic Gradient Boosting with Squared Error Epsilon Loss, 2014
