💼 Customer Churn Prediction & Analysis

🚀 Overview

This project analyzes and predicts customer churn using a real-world Telco dataset. We explore key factors influencing churn, build predictive models, and create an interactive Streamlit web app for churn risk scoring.

🎯 Business Goal: Identify customers likely to leave and suggest actionable strategies to improve retention.

---

💡 Key Features

• 📊 End-to-end analysis: From data cleaning and EDA to model building and business insights.
• 🤖 Advanced models: Logistic Regression, Random Forest, and XGBoost compared and interpreted.
• 💥 Interactive Streamlit app: Predict churn risk dynamically based on customer features.
• 📈 Business recommendations: Data-backed insights for churn reduction strategies.

---

🗂️ Dataset

• Telco Customer Churn dataset: Download here
• Contains customer demographics, services, billing, and churn labels.

---

⚙️ Tech Stack

• Python (pandas, scikit-learn, XGBoost, matplotlib, seaborn)
• Streamlit for web app
• Jupyter Notebook for EDA and modeling
• Git & GitHub Codespaces for development

---

💬 Business Insights

✔️ Month-to-month contract customers have high churn risk — suggest loyalty programs.
✔️ Fiber optic customers churn more — consider special retention offers.
✔️ Customers without security or tech support services are more likely to churn — offer bundled services.
✔️ High monthly charges correlate with churn — provide personalized discount or value packages.

---

📁 Folder Structure

.
├── customer_churn.ipynb # Notebook with EDA and modeling
├── app/
│ ├── streamlit_app.py # Streamlit web app
│ └── model/
│   ├── churn_model.json # Saved XGBoost model
│   └── feature_names.json # Model feature names
├── data/
│ └── WA_Fn-UseC_-Telco-Customer-Churn.csv
├── requirements.txt
├── README.md

---

🌟 Streamlit App

👉 Try it live: Streamlit App Link

Or run locally:

pip install -r requirements.txt
streamlit run app/streamlit_app.py

Made with love ❤️