🔍 Customer Churn Prediction using Machine Learning

---

📌 Project Overview

Customer churn is one of the most critical challenges faced by subscription-based businesses.
This project focuses on predicting customer churn using multiple machine learning classification algorithms, helping businesses identify at-risk customers early and take proactive retention actions.

🔹 The project includes data preprocessing, exploratory data analysis, model training, evaluation, and comparison of multiple ML algorithms to select the best-performing model.

🔗 GitHub Repo: Link

🌐 Live Demo: Link

---

🎯 Objectives

• 🔍 Analyze customer behavior data
• 🤖 Train and compare multiple ML classification models
• 📊 Evaluate models using Accuracy, Precision, Recall & F1-score
• 🏆 Select the best-performing model for churn prediction
• 🚀 Prepare the project for real-world deployment (Streamlit-ready)

---

🧠 Machine Learning Algorithms Used

• Logistic Regression
• Decision Tree Classifier
• K-Nearest Neighbors (KNN)
• Support Vector Machine (SVM)
• Naive Bayes
• Random Forest Classifier
• AdaBoost Classifier
• Gradient Boosting Classifier

---

📊 Model Evaluation Metrics

Each model is evaluated using:

• ✅ Accuracy
• 🎯 Precision
• 🔁 Recall
• 📈 F1-Score

A comparative performance table is generated to identify the best model.

---

🏆 Final Outcome

✅ XGBoost Classifier (trained separately) was identified as the best-performing model, delivering superior performance across all evaluation metrics.

📌 PCA was also tested, but no significant improvement in performance was observed.

---

📁 Project Structure

Customer-Churn-Prediction/
│
├── Customer Churn Prediction.ipynb   # Model training & analysis
├── app.py                            # Streamlit web application
├── customer_churn_classifier.pkl     # Trained ML model
├── customer_churn_large_dataset.xlsx # Dataset
├── README.md                         # Project documentation
├── requirements.txt                  # Python dependencies
├── .gitignore                        # Ignored files
└── venv/                             # Virtual environment (ignored)

---

⚙️ Tech Stack & Tools

• 🐍 Python
• 📊 Pandas, NumPy
• 🤖 Scikit-learn
• 📉 Matplotlib, Seaborn
• 🌐 Streamlit
• 📒 Jupyter Notebook

---

🚀 How to Run the Project

# Clone the repository
git clone https://github.com/your-username/customer-churn-prediction.git

# Navigate to project folder
cd customer-churn-prediction

# Install dependencies
pip install -r requirements.txt

# Open Jupyter Notebook
jupyter notebook

🖥️ Run Streamlit Web App

Run the Streamlit application

streamlit run app.py

Once the app starts, open the browser and visit: 👉 http://localhost:8501/

---

The web app allows:

• 📊 Viewing data insights
• 🤖 Understanding model performance
• 🔮 Predicting customer churn in real time

---

🌐 Deployment

This project includes a Streamlit-based web application and is ready for deployment on:

• Streamlit Cloud
• Localhost
• Any cloud VM (AWS / Azure / GCP)

The trained model (.pkl) is loaded directly into the app for real-time predictions.

📌 Key Learnings

• End-to-end ML pipeline development
• Handling categorical & numerical data
• Model comparison & evaluation
• Overfitting analysis
• Real-world ML project structuring

---

📄 License

This project is licensed under the MIT License.

---

👨‍💻 Author

Piyush Kumar
🚀 Data Science & Machine Learning Developer

📬 Let's Connect

🌐 Portfolio  •  💻 GitHub  •  💼 LinkedIn  •  📧 Email