Bike Sharing Demand Forecasting

Problem Statement

Predict daily bike rental demand for a bike-sharing system using historical usage data, calendar features, and weather conditions. The goal is to support 30-day ahead planning for fleet sizing and capacity management.

Dataset

• Source: UCI Bike Sharing Dataset
• Daily rental records from January 2011 to December 2012 (731 days)

Approach

Random Forest Regressor with feature engineering:

• Hyperparameter tuning: GridSearchCV with TimeSeriesSplit (8 folds)

Project Main Files

solution_madhu_tangudu.ipynb   # Main notebook (Parts 1–4)
utility.py                     # Helper functions and model pipeline
requirements.txt               # Python dependencies
data/input/day.csv             # Dataset

How to Run

# Install dependencies
pip install -r requirements.txt

# Launch notebook
jupyter notebook solution_madhu_tangudu.ipynb