/ projects

Hosting the DC Michelin Guide Data Science Challenge

Motivated by the Michelin Guide's 2016 inaugural DC Guide Book, I hosted a data science challenge to to most accurately predict which restaurants would make the cut. The Michelin Guide and General Assembly sponsored the competition, giving the winner $500 and two tickets to the French Ambassador's residence for the Guide's unveiling. Of the 30 entries received, three from the top five were from my students at General Assembly. Read more about it in Washington City Paper and Technical.ly DC. Check out the full competition rules and view submissions on my Github.


Presenting at the 2016 Boston Hockey Analytics Conference

Ice hockey is experiencing its "Moneyball Moment," where statisticians are identifying innovative methods to provide a competitive advantage. Brian Carothers and I were accepted to present our research at the 2016 Boston Hockey Analytics Conference. We argued for the recognition of a new defensive metric called "Recovery," demonstrating one's ability to remedy their errors. View our slides and code.


Measuring the Impact of Airbnb on Los Angeles County Rent Prices

Since 2008, Airbnb has seen exponential growth in its home-sharing platform, boasting over two million listings across 190 countries. Various cities have argued the platform causes homeowners to decrease the long-term tenant housing supply as owners substitute space for short-term tourist stays. Santa Monica, CA was the first municipality to explicitly ban stays shorter than 30 days. I analyzed the impact of of 22,000 Airbnb listings between January 2014 to December 2015 on 52 LA zip code rent prices. (A Medium blog post documenting results is forthcoming.)


Building a Wine Recommendation Engine + Data Education DC Meetup Presentation

The French Embassy sponsors the annual France On Campus Award, which encourages the spread of French culture on university campuses. I earned support on behalf of GW DATA, a data science student organization I helped create during my undergrad. We held a hackathon, and my four-person team built a beta wine recommendation engine. I learned so much about collaborative and content-based filtering that I subsequently spoke at a Data Education DC Meetup: Building a Wine Recommendation Engine (When You Don't Know Anything About Wine). Read about it in Technical.ly DC.