🧠 Customer Retention Dashboard

This is a real-world, end-to-end Data Analytics + Machine Learning project built by Banothu Prashanth to analyze customer behavior and predict churn using RFM segmentation, logistic regression, and a Streamlit dashboard.

---

📊 Project Highlights

✅ RFM Segmentation
✅ Churn Prediction Model (ROC AUC: 0.999)
✅ Interactive Dashboard (Streamlit)
✅ Cohort & Visual Analysis
✅ Industry-relevant folder structure
✅ Production-ready code and deployment

---

🚀 Features

• 📌 Clean raw customer data (data/raw/)
• 🧹 Processed using Pandas, visualized with Matplotlib/Seaborn
• 📈 RFM analysis to segment users (Loyal, At Risk, Churned)
• 🤖 Logistic Regression model built with Scikit-learn
• 📊 Live Streamlit dashboard to explore predictions
• 🧠 Stored ML model as .pkl for deployment

🧱 Folder Structure

customer_retention_dashboard/ │ ├── data/ │ ├── raw/ # Raw input data │ ├── processed_churn_data.csv # Final RFM dataset │ ├── churn_model.pkl # Trained ML model │ ├── *.png # Distribution plots │ ├── scripts/ │ ├── data_cleaning.py │ ├── rfm_analysis.py │ ├── churn_model.py │ ├── cohort_analysis.py │ ├── streamlit_app/ │ └── app.py # Streamlit dashboard │ ├── requirements.txt ├── .gitignore ├── README.md

🛠️ Tech Stack

• Python (Pandas, NumPy, Scikit-learn)
• Data Visualization: Matplotlib, Seaborn, Altair
• App: Streamlit
• Version Control: Git + GitHub

📈 Model Performance

• Model: Logistic Regression
• ROC AUC Score: 0.99904 🔥
• Precision: 0.99+
• Accuracy: 99%

🧠 Author

Banothu Prashanth
📧 banothuprashanth121@gmail.com
🌐 GitHub
🔗 LinkedIn: https://www.linkedin.com/in/banothu-prashanth-4406b3233