Crowdify is a web app that intelligently composes playlists on Spotify. An organiser can create an event, which guests attend through an URL in 1 click.
We then aggregate the music preferences of all the attendees, and generate a playlist which matches the target demographic as closely as possible. Playlist is automatically generated and updated on Spotify.
Built on Node/Express/Mongo during HackCambridge Recurse.
A secure Peer-to-Peer payment system leveraging accelerometer and gyroscope sensors in mobile devices. Unique identifier is generated per transaction by shaking 2 phones simultaneously while held together. Pairing of sender & receiver is done automatically on server side through digital signal processing of spatial data from devices.
Android mobile client & Python + Flask + Statsmodels backend for signal processing. Built in 24 hours at code::XtremeApps:: 2016 hackathon.
A project focused on improving user experience, we have created a personalized and innovative system for product recommendations.
We augmented Dixons Carphone’s products database with metadata scraped from external sources based on identifying product and manufacturer information. With our metadata, we are able to provide accurate recommendations of laptops to customers.
We are able to connect in-store and online shopping experience for customers, providing both the individually personalized support of a sales rep and convenience and ease of use of online shopping.
A machine learning project that predicts whether a project submitted to a hackathon is a winning project. Our classifier was trained on data scraped from over 30,000 projects made at 1000+ hackathons on DevPost. Testing revealed accuracy of around 95%, with good precision and almost perfect recall.
A tool written in haskell that draws Mandelbrot fractals. It is capable of generating stunning high resolution renders, both stills and animated, of the Mandelbrot set fractal featuring smooth colorful gradients.
Capable of leveraging multiple cores for rendering, and highly customizable output (see docs).
A project with the purpose on improving user experience and interface design. We created a search tool for checking HSA approved medications and devices in Singapore, while detecting illegal drugs and recommending approved alternaltives. Also includes pharmacy locator feature.
Research project undertaken at IHPC over period of 6 months aiming to analyse public sentiments using Twitter social media platform. I wrote a program to automatically collect tweets of interest, then annotated the tweets, trained and compared differernt machine learning classifiers and performed parameter optimization along with novel emoji based sentiment classification.
Data crunching was done and I made a web application to visualize the collected data. Presented and won bronze award (top in category). at Singapore Science and Engineering Fair 2014.