This repository contains one of my first submissions to the Kaggle Bike Sharing Demand competition.
The included notebook covers:
- Data cleaning and exploratory analysis
- Uncovered domain knowledge and its application to feature engineering
- Why the Gradient Boosting Regression Trees (GBRT) model was suitable for this problem
- Model training and evaluation (including hyperparameter tuning)
- Model scoring, areas for improvement, and submission