📊 Customer Churn Analysis

Using machine learning to predict which customers are likely to leave.

---

📝 Project Overview

Analyzes telecommunications customer data to understand churn patterns and build a predictive model.

• 7,043 customers with service, contract, and billing data
• 26.5% churn rate identified
• 80% model accuracy achieved with Random Forest
• $139,000+ monthly revenue loss from churned customers

📊 Key Findings

Contract Types (Biggest Factor)

• Month-to-month: 42.7% churn rate
• One year: 11.3% churn rate
• Two year: 2.8% churn rate

Payment Methods

• Electronic check: 45.3% churn rate (highest risk)
• Credit card: 15.2% churn rate
• Bank transfer: 16.2% churn rate

Customer Demographics

• Senior citizens: 41.7% churn rate
• Customers without partners: 32.9% churn rate
• New customers: highest risk group

🤖 Machine Learning Results

• Model: Random Forest Classifier
• Accuracy: 80%
• AUC Score: 0.8406
• High-risk customers identified: 109 customers
• Top predictive factors: Tenure, Total Charges, Monthly Charges, Contract Type

💡 Recommendations

1. Focus on contracts: Encourage longer-term contracts
2. Fix payment issues: Migrate electronic check users to automatic payments
3. Improve fiber service: Address quality problems (30.9% churn rate)
4. Support new customers: Extra attention in first few months
5. Target high-risk customers: Use ML model for proactive retention

🚀 Quick Start

1. Setup environment:

   python3 -m venv .venv
   source .venv/bin/activate
   pip install -r requirements.txt

2. Run analysis:

   jupyter notebook Churn_analysis.ipynb

3. View results:

   • Charts saved to figures/ folder
   • Model identifies high-risk customers
   • Financial impact calculated

📁 Files

• Churn_analysis.ipynb - Main analysis notebook
• Customer-Churn.csv - Dataset (7,043 records)
• requirements.txt - Dependencies (pandas, numpy, matplotlib, sklearn)
• figures/ - Generated visualizations

---

Competition Entry: Data Analytics with AI – Contest #1 | Date: September 30, 2025