📞🔍 Vishing Detection System

Django Backend + Android Frontend

This project is a Vishing (Voice Phishing) Detection System that uses an Android mobile application to intercept incoming phone calls and send call information to a Django backend for real-time analysis. The backend analyzes the call data and determines whether the call is safe, suspicious, or potentially harmful.

---

🚀 Overview

This system consists of two major components:

1. Android App (Frontend)

• Runs on the user’s phone.
• Intercepts incoming calls (using Android Telephony API).
• Extracts caller phone number and metadata.
• Sends call information to the Django backend via REST API.
• Displays the backend’s analysis (Safe / Suspicious / Vishing).

2. Django Server (Backend)

• Receives call data from Android devices.
• Runs ML or rule-based vishing detection logic.
• Returns a JSON response with the risk analysis.
• Stores logs for future training and improvement.

---

🧠 How It Works (System Flow)

1. A call arrives on the Android phone.

2. Android App captures:

   • Phone number
   • Call time
   • Call type (incoming/outgoing)

3. App sends this data to Django backend → /api/call-detection/

4. Django backend:

   • Processes the phone number
   • Checks machine-learning or blacklist rules
   • Generates a "risk score" or "threat category"

5. Backend responds with:

   {
     "status": "suspicious",
     "score": 87,
     "reason": "Number reported multiple times"
   }

6. Phone displays an alert (RED / YELLOW / GREEN).

7. Django saves the call analysis for future dataset building.

---

🏗️ Technologies Used

Backend (Django)

• Django REST Framework
• Python
• SQLite / PostgreSQL
• Machine Learning model or rule-based detector

Frontend (Android)

• Android Studio (Java/Kotlin)
• TelephonyManager
• Retrofit for API calls
• Notifications / UI alerts

---

📦 Django API Endpoints

POST /api/call-detection/

Send call details to backend.

Example request:

{
  "phone_number": "+123456789",
  "call_type": "incoming",
  "timestamp": "2025-11-16T10:30:00"
}

Example response:

{
  "status": "vishing",
  "confidence": 0.92,
  "message": "Number associated with scam patterns"
}

---

🛠️ Setup Instructions (Backend)

1. Clone the project

git clone https://github.com/yourusername/vishing-detector-backend.git
cd vishing-detector-backend

2. Create virtual environment

python3 -m venv venv
source venv/bin/activate

3. Install dependencies

pip install -r requirements.txt

4. Apply migrations

python manage.py migrate

5. Run server

python manage.py runserver

---

📱 Android App Setup (Frontend)

1. Open /android_app/ folder with Android Studio

2. Update backend URL in Retrofit service:

   const val BASE_URL = "http://YOUR_SERVER_IP:8000/"

3. Run on device

4. App requests permissions:

   • READ_CALL_LOG
   • READ_PHONE_STATE
   • POST_NOTIFICATIONS

5. Intercepted calls will now be analyzed by Django server.

---

🤖 Detection Logic

You may use:

Option 1 — Machine Learning Model

• Train using past call logs + labels (safe/spam/vishing)
• Use logistic regression, SVM, or transformer language model on call remarks

Option 2 — Rule-Based Detection

• Check phone number:

  • Reported spam lists
  • Known fraud numbers
  • Short suspicious patterns (e.g., “+255 6XX 00XX”)
  • Time-based anomalies

---

📊 Database Example

Phone Number	Status	Score	Timestamp
+255621XXXX	vishing	0.91	2025-11-15
+1415555XXXX	safe	0.02	2025-11-16
+255784XXXX	suspicious	0.62	2025-11-16

---

📜 License