SMART TENDER FRAUD DETECTION SYSTEM USING GNN

• 

  Authentication

• 

  Dashboard

• 

  Dashboard

• 

  Create tender

• 

  history of tenders

• 

  Company Details

• 

  Pattern Analysis

• 

  Risk Analysis

• 

  Add Company

• 

  Alerts

• 

  Report Generation

• 

  Setting

Inspiration

What it does

How we built it

Challenges we ran into## About the Project

The Smart Tender Risk Detection System was inspired by the need to improve transparency and identify unusual bidding patterns in e-procurement processes. Traditional tender monitoring can make it difficult to recognize repeated relationships between companies, similar quoted prices, and changes in winning patterns across multiple tenders.

The system uses graph-based analysis and AI-assisted risk assessment to examine tender information, participating companies, quoted prices, and historical bidding relationships. It identifies suspicious patterns, generates a suspicion score, and provides explanations with supporting evidence to help procurement authorities decide when further investigation may be required.

The project was developed using React.js, Tailwind CSS, FastAPI, Python, and PostgreSQL. The system includes secure authentication, tender management, company and bid management, historical relationship analysis, and risk monitoring. The AI component is designed to use graph-based techniques, including Graph Neural Networks (GNNs), to represent relationships between companies and tenders.

One of the main challenges was connecting historical bidding records, calculating meaningful relationships, and ensuring that risk explanations were based on actual data rather than arbitrary scores. Another important consideration was handling insufficient historical data and presenting results responsibly.

Through this project, I gained practical experience in full-stack development, database design, API integration, data analysis, and AI-based risk assessment. The system is intended to support transparent procurement and manual investigation; a suspicious pattern or high risk score alone does not prove fraud.

Accomplishments that we're proud of

What we learned

What's next for SMART TENDER FRAUD DETECTION SYSTEM USING GNN

Built With

• fastapi
• graph
• postgresql
• python
• react.js
• tailwind-css