Student Project
Capital Bikeshare
Team members:
Our team created an Extra Trees Regressor to predict hourly demand for bikes used in the Capital Bikeshare system in Washington D.C. In addition to using the data provided by Kaggle on bikeshare usage, we incorporated third-party hourly precipitation data to boost our predictive accuracy. Our initial naive model ranked 1193rd, while our final submission ranked 250th, which put us within the top 15% at the time of submission.
Course
Data Science 207. Applied Machine Learning, Fall 2014
Kaggle Bikeshare