PROJECTS

Here are some projects of predictive models, interactive visuals, and cognitive science research that I have worked on recently.

DataVizCollection

DATA VIZ PROJECTS

I enjoy designing and creating data visuals - for clients, as challenge contributions (e.g. '#TidyTuesday), and in my spare time. These graphics here are generated with a code-first approach and almost all of them do not involve any manual post-processing steps. View TidyTuesday plots code repository

penguins

PENGUINS LEARNING DASHBOARD

Built with R flexdashboard and Shiny, this interactive web dashboard facilitates learning about different species of Antarctica penguins. Also, a demonstration of dynamic data visualisation of an open dataset. Launch Penguins Learning Dashboard

music

MUSIC MATCH MAKER

I developed a song analyzer and recommendation tool, based on Random Forest and tested with multiple other ML algorithms, using personal Spotify listening histories and song feature data acquired from Kaggle. View Music Match Maker on GitHub

curiosity_ggplots

MOTIVATED RISKY-TAKING

Can curiosity kill the cat? Using (generalised) mixed-effects modelling, structural equation modelling and Bayesian inference, this research project shows similar cognitive-neural mechanisms underlying the driving force of curisosity and extrinsic instrumental rewards for risk-taking (in a group of brain areas called the brain's reward networks). This work is published in the journal Nature Human Behaviour

plastic

OCEAN PLASTIC BAG OBJECT CLASSIFIER

A simple image classifer that identifies plastic bags in the ocean, built with FastAI using transfer learning (from pre-trained neural networks). Try Ocean Plastic Classifier

mvpa

BRAIN ACTIVITIES PREDICT GAMBLE DECISIONS

This project applied multivariate pattern analysis (incorporating Support Vector Classification & Multiple Kernel Learning ML algorithms) to decode brain activity patterns and predict gambling decisions. View MVPA analysis on GitHub

map

NEUROHACKMAP

My first Hackathon collaboration - In 2019 summer I partook in a two-week Neurohackademy with the University of Washington, in Seattle, and contributed to a code project which generates interactve visuals on geospatial data related to participants' career movements. View Neurohackmap project on GitHub

BCOS

BCoS & OCS COGNITIVE SCREEN BATTERIES

During my Masters and PhD study in Birmingham, I contributed to this line of large-scale clinical trials to develop and validate the assessment tools on post-stroke cognitive functions. Visit OCS-Cantonese project site

Read BCoS publication