I entered college as primarily a political science major, but decided to add computer science as a second major since I came in with a lot of credits and had a mild interest in programming. I enjoyed CS much more than I thought it would. It reminded me of chess, where success requires a mix of logic and creativity. I also felt I could have a much greater impact through software, because it scales so quickly. While I stayed with both majors, I gravitated towards CS because I knew technology had the greatest potential to make the world a better place.
A few of my projects are listed on GitHub, but below are a few of the more interesting ones.
Circa is the primary product of the startup I co-founded with Andrew Schuster. Circa is a platform that makes local buying and selling easy by handling the payments, delivery, and returns. Cheaper than Amazon, safer than Craigslist! What made Circa unique was our offering of last-mile delivery for $2 a package by utilizing on-the-way delivery; packages were paired with people who were already driving toward the packages’ destination. Ultimately, our mission was to improve communities’ standards of living and reduce their environmental impact by allowing them to earn an income from reusing goods their neighbors no longer need and would otherwise throw away.
I solely developed the entire initial version of Circa (the “minimum viable product”), and afterward Andrew and I worked on the codebase together while I ran the business end of things as CEO. While we achieved strong growth (on average, ~80% month-over-month during our active months), we didn’t hit profitability before running out of funds and weren’t able to raise any funds.
After the company shut down, I put Circa in test mode at www.usecirca.com and open sourced to code on GitHub.
ChessDB.xyz (In progress)
What’s the best episode of a given TV show? A simple “Hot or Not” style game that shows you two episodes of a TV show, and you choose the better one. This is a treated as a “match” between the two episodes and the Elo rating algorithm to update each episode’s rating and ranking. With enough people rating episodes, this system creates extremely accurate episode rankings, which can then be used to calculate all sorts of interesting stuff, like the best/worst seasons, best/worst episodes within a season, etc.
You can see a specific iteration of this at www.ratetheoffice.com.
I hadrecently taken up an interest in machine learning, and I wanted to implement a machine learning algorithm via a useful project. At the time I was suffering inbox overload, so I set about writing software that could automatically tell me whether an email was worth reading before I read it. The code was hacked together pretty quickly, but I was still able to classify emails as urgent/important with about 75-80 percent accuracy. Google has since unveiled their three tabbed Gmail classification system (primary, social, and promotional) which made the issue less pressing, so I’ve moved on to other projects.
First created: Mid-2013
Last updated: Mid-2013
Game Rating Calculator (Android)
My friends and I love to play the game Settlers of Catan. We’re pretty competitive about it, and would argue over who was the best player. One night I decided I’d build a desktop app to end the debate once and for all. I developed a rating algorithm based on the Elo rating formula used in chess, the main difference being that my algorithm takes the number of players into account. As of this writing, we’ve played over 100 rated games, and the ratings have turned out to be very accurate (I’m ranked first, so maybe I’m a bit biased). The app can be used for pretty much any game that has some amount of skill, which is why it was renamed from “Settlers of Catan Rating Calculator” to simply “Game Rating Calculator”.
Although it’s a simple Android app I made just for fun, I was pleasantly surprised: the app has over 1000 downloads, and usually gets a couple new downloads every day. I’ve even been asked to make an iOS version! While I wrote all the code for the initial version and developed the algorithm, my friend Andrew Schuster updated the code in 2014 so it’d look better on newer Android devices.
Code available on GitHub.
Current Version: 2.1.1
Last Updated: May 15, 2014
GPR Article Recommender (has since been removed, though not by me :( )
During my time at the Georgia Political Review, I noticed that we had a high bounce rate on our website. About 80 percent of readers would read one article and then leave the site. To keep readers more engaged, I suggested creating an article recommender that would place related articles at the end of each article. Since we don’t track users’ browsing habits, a traditional recommender system wouldn’t work. Instead, I’ve turned this into a classification project, where the recommender classifies each article into one of 10 preexisting categories and recommends the three most recent articles in that category.
Code on GitHub.
Game Rating Calculator (desktop)
I made this eons ago, when I first began programming on a somewhat regular basis. Alas, the source code has been lost to the sands of time.
Current Version: 2.0.3
Last updated: 8/11/12
Blog Entries Under “Software”
How to Eat Dinner with Barack Obama (February 4, 2013)
Building a Rating System (December 16, 2012)