Hotel Reservation Prediction

Project Goal:
Predict which customers are likely to cancel their reservation or checkout before check-in using historical hotel booking data. This enables hotels to take proactive actions to reduce revenue loss.

Table of Contents

• Project Overview
• Features
• Tech Stack
• Architecture & Pipeline
• Installation
• Usage
• Project Structure
• Contribution

Project Overview

The Hotel Reservation Prediction project is an end-to-end machine learning solution for predicting early checkouts or cancellations. The pipeline automates data ingestion, preprocessing, model training, and deployment using modern CI/CD practices. A Flask-based web app allows interactive prediction requests via Docker deployment.

Features

• Predict customer checkout behavior before check-in.
• ML model trained with LightGBM.
• Automated preprocessing and feature engineering.
• Experiment tracking with MLflow.
• Fully automated CI/CD pipeline using Jenkins, Docker, and Google Cloud Platform (GCP).
• Integration with GCP Buckets and Container Registry.
• Role-based access with IAMs and service accounts.
• Flask web app for serving predictions in real-time.

Tech Stack

• Programming: Python
• Data Processing & ML: Pandas, NumPy, LightGBM, Scikit-learn
• Web App: Flask
• CI/CD: Jenkins, Docker, GitHub
• Cloud: Google Cloud Platform (GCP) – Storage Buckets, Container Registry
• Experiment Tracking: MLflow
• Version Control: Git & GitHub

Architecture & Pipeline

1. Data Ingestion: Raw hotel booking data stored in GCP buckets is ingested via Python scripts.
2. Data Processing: Cleaning, preprocessing, and feature engineering applied to prepare training datasets.
3. Model Training: LightGBM model is trained, evaluated, and saved to artifacts/models.
4. MLflow Tracking: All experiments, metrics, and models are tracked for reproducibility.
5. Deployment:
   • Flask app serves predictions.
   • Docker image built and pushed to GCP Container Registry.
   • Deployed with full CI/CD pipeline using Jenkins.
6. Automation & Security:
   • Jenkins pipelines handle automated builds, tests, and deployments.
   • IAM roles and service accounts secure GCP resources.

Installation

1. Clone the repository:

git clone https://github.com/Subrat1920/Hotel-Reservation-Cancellation-Detection-MLOps.git
cd Hotel-Reservation-Prediction

2. Install dependencies:

pip install -r requirements.txt

3. Set environment variables for GCP authentication (if using service accounts):

export GOOGLE_APPLICATION_CREDENTIALS="path/to/service-account.json"

Usage

A. Local Execution (Training & Prediction)

1. Start the Flask app:

python app.py

2. Access the app at http://localhost:5000 to make predictions.

B. Dockerized Deployment

1. Build the Docker image:

docker build -t hotel-reservation-prediction .

2. Run the container:

docker run -p 8080:8080 hotel-reservation-prediction

3. Access at http://localhost:8080.

4. CI/CD Pipeline
   Jenkinsfile handles automated cloning, testing, building, pushing Docker image to GCP, and deploying.

Project Structure

├── 📁 artifacts/
│   ├── 📁 models/
│   │   └── 📄 lgbm_model.pkl
│   ├── 📁 processed/
│   │   ├── 📄 process_test.csv
│   │   └── 📄 process_train.csv
│   └── 📁 raw/
│       ├── 📄 raw.csv
│       ├── 📄 test.csv
│       └── 📄 train.csv
├── 📁 config/
│   ├── 🐍 __init__.py
│   ├── ⚙️ config.yaml
│   ├── 🐍 model_params.py
│   └── 🐍 path_config.py
├── 📁 custom_jenkins/
│   └── 🐳 Dockerfile
├── 📁 pipeline/
│   ├── 🐍 __init__.py
│   └── 🐍 training_pipeline.py
├── 📁 src/
│   ├── 🐍 __init__.py
│   ├── 🐍 data_ingestion.py
│   ├── 🐍 data_processing.py
│   ├── 🐍 exception.py
│   ├── 🐍 logger.py
│   └── 🐍 model_training.py
├── 📁 static/
│   └── 🎨 style.css
├── 📁 templates/
│   └── 🌐 index.html
├── 📁 utils/
│   ├── 🐍 __init__.py
│   └── 🐍 common_functions.py
├── 🚫 .gitignore
├── 🐳 Dockerfile
├── 📄 Jenkinsfile
├── 📖 README.md
├── 🐍 app.py
├── 📄 requirements.txt
└── 🐍 setup.py

Contributing

Fork the repository.
Create a feature branch: git checkout -b feature-name
Commit your changes: git commit -m "Add feature"
Push to branch: git push origin feature-name
Create a Pull Request.