Churn Prediction Using Machine Learning

This project focuses on predicting customer churn using various machine learning models, with an added business-facing Power BI dashboard for visual insights. The dataset is based on a real-world telecom provider and includes customer demographics, account information, and service usage patterns.

The goal was to not just build models, but to also understand what factors drive churn, and how we can visualize risk in a way that non-technical users can act on.

---

📁 Project Structure

+---data
|   |__ Telco-Customer-Churn.csv
|
+---notebooks
|   |__ Churn_prediction.ipynb
|
+---outputs
|   |__ churn_cleaned.csv
|   |__ Churn_prediction_BI_Final.pbix
|
+---Images
    |__ dashboard.png
    |__ roc_curve.png

🚀 Key Highlights

• 🔍 Exploratory Data Analysis using Seaborn and Matplotlib
• 🔄 Data Preprocessing including encoding, scaling, and null handling
• 📊 ML Models: Random Forest, SVM, XGBoost
• 📈 Evaluation: ROC-AUC, F1, confusion matrices
• 📊 Power BI Dashboard: Highlights churn risk, service patterns, and user segments

🔧 Tools & Libraries

• Python, Pandas, NumPy
• Scikit-learn, XGBoost, SHAP
• Seaborn, Matplotlib
• Power BI (for dashboard)
• Git/GitHub for version control

📈 Sample Results

Model	Accuracy	F1 Score	ROC-AUC
Random Forest	78%	0.55	0.823
SVM	79%	0.55	0.80
XGBoost	77%	0.55	0.822

📊 Power BI Dashboard Preview

The dashboard provides a business-friendly summary:

• Churn by tenure, gender, and contract type
• Service subscription trends
• Risk distribution across customer segments

---

▶️ How to Run

1. Clone the repository
2. Install dependencies:

pip install -r requirements.txt