DataFest2017 Expedia Challenge Apr, 2017
Our approach focused on recognizing the patterns for Expedia Mobile users. We noticed that there is a noticeable increase in bookings under 200 miles for mobile users compare to non-mobile users. We realized that these users might not be traveling by air, but instead, by car. Therefore, it proves the point that people traveling by car with shorter distance are more likely to book a hotel right before their trip, they are also more likely to use mobile phone due to computer availability in these situations. There is also the possibly of short-term business trip. We then figured that how mobile users tend to have shorter traveling distance and shorter length of stay.
American Statistics Association DataFest 2017: Best Insight Award, 1st place Emory site.
Predicting Blood Donation Sep, 2016 – Nov, 2016
This is my first Data Science competition. Surprising, simple Logistic Regression turns out to perform really well compared to k-NN, random forest, etc due to the small size of the dataset.
Top 3% at the time of last submission(Nov 2016).
House Price Prediction Jan, 2017
Santander Satisfaction Challenge Apr, 2017
Final project for CS378 course at Emory. Our Proposal.