Beat The Crowd

Sunday. February 11, 2018 - 1 min

Project completed across 2 weeks as a Data Science Fellow at Insight Data Science. Uses historical passenger trip data from BART and weather history data from Weather Underground to predict crowd levels on BART with high accuracy (93% of variance explained in test data). Passenger volume prediction driven by a random forest algorithm.

Deployed as an interactive front-end interface via Flask.

All code written in Python using the following tools (among others):

pandas
scikit-learn
Matplotlib
seaborn
lxml
flask

Check out the GitHub repo to see the code for this project.