Bike Rental Demand Prediction Project

This project predicts bike rental demand using machine learning techniques. It downloads data from a specified URL, performs feature engineering, trains several regression models (including a stacked model), and generates a submission file.

Table of Contents

1. Introduction
2. Features
3. Getting Started
   • Prerequisites
   • Installation
   • Usage
4. Project Structure
5. Model Training
6. File Descriptions
7. Contributing
8. License

Introduction

This project aims to build a predictive model for bike rental demand based on various environmental and temporal features. It demonstrates a full data science workflow from downloading data to making predictions using various models.

Features

• Data Download: Downloads compressed data from a specified URL.
• Data Preprocessing: Handles missing values, performs feature engineering, and encodes categorical features.
• Model Training: Implements several regression models:
  • Ordinary Least Squares (OLS) Regression
  • Decision Tree Regression
  • Random Forest Regression
  • AdaBoost Regression
  • XGBoost Regression
  • Stacked Model
• Model Evaluation: All models are validated using the RMSLE error metric.
• Feature Importance: Provides a visualization of XGBoost feature importances.
• Submissions: Generates a submission CSV file with predictions.
• Command-Line Interface: Uses argparse to allow configuration of paths from the command line.
• Logging: Provides informative logging of project execution.

Getting Started

Prerequisites

• Python 3.6 or higher
• pip (Python package installer)
• Internet connection (for data download)

Installation

1. Clone the repository:

   git clone https://github.com/your-username/your-repo-name.git
   cd your-repo-name

2. Install the required packages:

   pip install -r requirements.txt

Usage

1. Download the data and generate a submission file

   python bike_rental_project.py --input_path /path/to/your/input --working_path /path/to/your/working --submission_file submission.csv

   • Replace /path/to/your/input with the desired input directory, /path/to/your/working with the working directory, and submission.csv with the desired name of the submission file.

   or for the default folders:

   python bike_rental_project.py

• The model will use the /kaggle/input and /kaggle/working folders for input, and create the output submission.csv file, you can use the parameter --force_download if you want to download again the data.

Project Structure

├── bike_rental_project.py # Main Python script ├── requirements.txt # Libraries to install └── README.md # This file

Model Training

The code trains several models, and it shows the feature importances of the XGBoost Model, and saves the submission of the stacked model.

File Descriptions

• bike_rental_project.py: Contains the main Python script with all the project logic.
• requirements.txt: A list of the python packages required to execute the program.
• data: The input folder, where the training and test set reside.
• submission.csv: the submission that was created

Contributing

Contributions are welcome! To contribute:

1. Fork the repository.
2. Create a new branch for your feature or bug fix.
3. Make your changes and commit them.
4. Push your changes to your fork.
5. Submit a pull request.