Detecting "Zombie Subscriptions" through AI-driven recurring transaction analysis and behavioral modeling.
Sub-Zero is a full-stack fintech web application that helps users identify Zombie Subscriptions—recurring payments that continue despite little or no meaningful usage.
Instead of simply displaying bank transactions, Sub-Zero analyzes recurring payment patterns, estimates financial leakage, generates risk scores, and provides AI-powered financial insights.
The project simulates modern fintech architecture inspired by India's Account Aggregator (AA) ecosystem using synthetic banking transaction data.
- 🔍 Recurring subscription detection
- 🧠 Zombie subscription risk scoring
- 📊 Interactive financial dashboard
- 💬 AI Financial Copilot
- 📈 Wealth projection calculator
- 📄 CIBIL-style Subscription Audit Report (PDF)
- 🔄 Simulated UPI AutoPay revocation workflow
- 🏦 Simulated Account Aggregator consent flow
User
│
▼
HTML • CSS • JavaScript
│
▼
Flask Backend
│
┌─────────┼─────────┐
▼ ▼ ▼
Analytics AI Copilot Reports
│
▼
Recurring Detection
│
▼
Zombie Risk Engine
│
▼
SQLite / Synthetic Data
- HTML5
- CSS3
- Bootstrap
- JavaScript
- Chart.js
- Python
- Flask
- Pandas
- NumPy
- ReportLab
- Git
- GitHub
- Gunicorn
- Render
The user connects their account through a simulated Account Aggregator consent flow.
Synthetic banking transactions are loaded.
Transactions are grouped by merchant.
Intervals between recurring payments are calculated.
Example:
Netflix
5 Jan
↓
4 Feb
↓
6 Mar
≈ 30 Days
The system checks usage activity.
If:
- recurring payment exists
- usage is very low
↓
Subscription is flagged as a Zombie Subscription
A risk score is assigned.
Financial insights, wealth projections, and audit reports are generated.
Sub-Zero
│
├── app.py
├── requirements.txt
│
├── analytics/
├── reports/
├── static/
├── templates/
├── data/
└── README.md
- Real Account Aggregator integration
- OpenAI-powered financial advisor
- SQLite persistent backend
- Multi-user authentication
- Fraud detection engine
- Spending heatmaps
- Subscription forecasting
(Add dashboard screenshots here)
https://sub-zero-8cft.onrender.com
Neha Kulkarni
GitHub: https://github.com/Nehakulkarni31
LinkedIn: (Add your LinkedIn profile)