Check out my latest open source projects below.

UIUC Bus Tracker

Thousands of students at U of I rely on the Champaign-Urbana buses each day as a means of getting to class. Unfortunately, many apps to check bus times take too long to load. As college students, we need this information as soon as possible. This is a beautiful, minimalist web app with load times under one second. No clutter, no hassle, just info. Check it out on GitHub.

Facebook Archive Parser

Facebook offers a neat feature which allows a user to download all their data as an archive file. This includes all the (possibly thousands) of messages they've ever sent or received. Wouldn't it be cool to do analytics, machine learning, etc. with this data? Within seconds, this Python wrapper allows users to generate a word cloud with this data and access a neatly-organized list of every message.

Twitter Support Vector Machine

Recognized by the American Statistical Association, this Python script uses the power of machine learning to do complex authorship attribution on thousands of tweets. After acquiring an Twitter API key, user can input as many public channels as they would like, and the script loads and parses every tweet, generates confusion matrices, finds most distinct words, and returns a trained and cross-validated classifier. In classifying two channels, this is ~95% accurate!

Want to learn more?

Email me at josh@joshbyster.com. I'd love to hear from you.