Bike Sharing Demand Prediction 🚴‍♂️

This project predicts bike rental demand using the Kaggle Bike Sharing Demand dataset.
It leverages AutoGluon TabularPredictor to build and optimize models automatically.

📊 Project Workflow

1. Dataset Preparation

   • Data obtained from Kaggle.
   • Split into train, test, and sample submission files.

2. Baseline Model

   • Trained a simple AutoGluon model on the raw training data.

3. Feature Engineering

   • Extracted features from the datetime column (hour, day, month).
   • Improved model accuracy with enriched features.

4. Hyperparameter Tuning

   • Adjusted AutoGluon model hyperparameters.
   • Compared results across training runs.

5. Evaluation

   • Measured prediction accuracy at each stage.
   • Demonstrated improvement from baseline → feature engineering → tuned model.

🛠️ Tech Stack

• Python
• Pandas
• AutoGluon
• Jupyter Notebook
• Sagemaker Studio