✈️ Flight Fare Finder

End-to-End Machine Learning Pipeline 🚀

---

---

📚 Project Overview

This project is a full MLOps-grade system that predicts flight fare prices based on flight details such as airline, route, booking class, and time to departure.

• ⛓️ Modular pipeline stages
• 🧪 ML experiment tracking with MLflow
• 📦 Data & model versioning with DVC
• 🚀 REST API with FastAPI
• 🐳 Docker containerization
• ⚙️ CI/CD automation via GitHub Actions & AWS ECR

---

🏧 Project Architecture

Data Ingestion → Data Validation → Data Cleaning →
Data Transformation → Model Training → Model Evaluation →
Prediction API →Docker Containerization → CI/CD → AWS ECR Deployment

---

🚀 Tech Stack

Category	Tools Used
Language	Python 3.10
ML Framework	Scikit-learn, XGBoost
Experimentation	MLflow + Dagshub
Versioning	DVC
Deployment	FastAPI + Uvicorn
Packaging	Docker
Automation	GitHub Actions → AWS ECR → EC2 Deployment

---

📂 Folder Structure

├── config/
├── src/mlproject/
│   ├── components/
│   ├── pipelines/
│   ├── config/
│   ├── entities/
│   └── utils/
├── artifacts/
├── dvc.yaml
├── params.yaml
├── schema.yaml
├── Dockerfile
├── app.py
├── README.md
├── requirements.txt
└── .github/workflows/cicd.yml

---

💠 Setup Instructions

🔹 Clone the repository

git clone https://github.com/JavithNaseem-J/FareFinder.git
cd FareFinder

🔹 Create & activate virtual environment

conda create <env name> python=3.10 -y
conda activate <env name>

🔹 Install dependencies

pip install -r requirements.txt

---

🧪 Running the Pipeline (via DVC)

Each pipeline stage is DVC-tracked and reproducible.

Stage	Command
Data Ingestion	python main.py --stage data_ingestion
Data Validation	python main.py --stage data_validation
Data Cleaning	python main.py --stage data_cleaning
Data Transformation	python main.py --stage data_transformation
Model Training	python main.py --stage model_training
Model Evaluation	python main.py --stage model_evaluation

Run the full pipeline:

dvc repro

---

📈 MLflow Tracking (via Dagshub)

• Logs parameters, metrics (R², MAE, MSE), models
• Stores all experiments and best model in the MLflow registry

---

🐳 Docker Support

Build Docker image:

docker build -t flight-fare-app .

Run the container:

docker run -p 8080:8080 flight-fare-app

---

⚙️ CI/CD Pipeline (GitHub Actions + AWS ECR)

Your CI/CD workflow includes:

• ✅ Code linting
• ✅ Unit tests (placeholder)
• ✅ Docker image build
• ✅ Image push to AWS ECR
• ✅ Auto-deploy to EC2 container (self-hosted)

Workflow file:

.github/workflows/cicd.yml

---

🧠 Key Highlights

• ✅ End-to-End ML lifecycle pipeline
• ✅ Model tuning via RandomizedSearchCV
• ✅ MLflow-based experiment tracking
• ✅ CI/CD auto-deployment with GitHub → AWS
• ✅ Production-grade FastAPI backend

---

📄 License

Distributed under the MIT License.
See LICENSE for more information.