Retention AI - Customer Churn Prediction System (Django + Ensemble AI)

Retention AI is an AI-powered web platform that helps businesses predict customer churn with high accuracy. Using an ensemble of advanced machine learning models — CatBoost, LightGBM, and XGBoost — it identifies customers who are likely to leave and provides intelligent recommendations to retain them.

🧠 Built with Django, RetentionAI allows:

• 📋 Single customer predictions through a smart form.
• 📁 Bulk churn analysis via CSV uploads.
• 🔍 AI-driven insights and retention suggestions based on customer behavior.

Ideal for telecoms, subscription platforms, and service-based businesses seeking to reduce customer loss and improve decision-making.

---

✨ Features

• ✅ User Authentication (Signup/Login)
• ✅ Single Customer Prediction (via smart form)
• ✅ Batch Prediction (upload CSV file)
• ✅ AI Reasoning & Suggestions (based on churn factors)
• ✅ Admin Dashboard and User Profile Views
• ✅ Clean, modern UI with loading animations & validation

---

🎓 Tech Stack

Layer	Tech
Backend	Django
ML Models	CatBoost, LightGBM, XGBoost
Frontend	HTML, CSS, Bootstrap Icons
Visualization	None (lightweight UI)
Auth	Django Auth

---

⚡ Setup Instructions

1. Clone the Project

git clone https://github.com/yourname/churn-prediction-django.git
cd churn-prediction-django

2. Create & Activate Virtual Environment

python -m venv env
source env/bin/activate      # On Windows: env\Scripts\activate

3. Install Requirements

pip install -r requirements.txt

4. Prepare the Models (Already Trained)

Place the following 6 files in:

churn/models/
├── catboost_model.cbm
├── lgb_model.pkl
├── xgb_model.pkl
├── best_threshold.txt
├── cat_columns.json
└── onehot_columns.json

5. Run Migrations & Server

python manage.py migrate
python manage.py runserver

---

⚙ How It Works

• Form input is parsed and encoded in two formats:

  • Raw (for CatBoost)
  • One-hot (for LightGBM/XGBoost)

• All 3 models return probabilities, which are averaged.

• If probability > threshold (optimized during training), system flags churn.

• Reasons/suggestions are generated based on business logic.

---

📚 Sample Use Case

A telecom company wants to predict customer churn:

• A support agent logs in.
• Enters customer details via form.
• Gets churn probability + AI suggestions (e.g. reduce pricing).
• Or uploads CSV for 1000 customers in bulk.

---

📚 Sample CSV Format

tenure,MonthlyCharges,TotalCharges,gender,SeniorCitizen,Partner,Dependents,PhoneService,...
24,75.5,1600.2,Female,0,Yes,No,Yes,...

(Include all model-required columns)

---

🖼️ Screenshots

🏠 Home & Landing Page

Home Page	Actions	Landing

---

🔐 Authentication Pages

Sign In	Sign Up

---

🧠 Prediction (Single)

Question Sample 1	Question Sample 2	Result

---

📊 Batch Prediction

Upload CSV	CSV Uploaded	Batch Result

---

🛠️ Admin Dashboard

Dashboard View 1	Dashboard View 2	Dashboard View 3

---

💼 Services

Services	Services (Alt)

---

🌟 Credits

• ML Modeling by: SyedSaadAli
• Django Development by: SyedSaadAli
• UI/UX: Bootstrap 5, FontAwesome

---

✉ Contact

For queries, suggestions, or improvements, feel free to reach out:

📧 syedsaadi427@gmail.com