67Frauders

Login page
Dashboard
History
Swipe and Review
Downloading file in process

Inspiration

Fraud detection is often presented as a machine learning problem, but the challenge showed us that finding fraud is only half the battle. Human reviewers still need to understand why a transaction was flagged and decide what action to take.

We wanted to build a platform that does more than detect fraud. We wanted to create a tool that helps reviewers investigate suspicious activity quickly, understand fraud patterns, and make confident decisions.

Our goal was to make fraud analysis feel less like reading a spreadsheet and more like solving a case.

What it does

67Frauders is an interactive fraud investigation platform that analyzes transaction data and helps reviewers identify suspicious activity.

The platform:

Uploads and analyzes transaction data
Flags suspicious transactions with risk scores
Explains why each transaction was flagged
Supports review actions such as Approve, Reject, and Escalate
Provides a keyboard-friendly review workflow
Generates investigation summaries
Visualizes fraud relationships through the Fraud Constellation network
Helps reviewers discover connections between cards, devices, IP addresses, merchants, and transactions

The result is a faster and more intuitive fraud review experience.

How we built it

Frontend

HTML
Interactive dashboards and investigation views

Backend [ADD BACKEND DETAILS HERE]*******************************

Key Features

Fraud review queue
Explainable risk scoring
Cost threshold simulator
Investigation notes
Audit workflow
Fraud Constellation visualization

The platform was designed to feel like a real fraud operations tool rather than a simple analytics dashboard.

Challenges we ran into

One of our biggest challenges was balancing fraud detection with reviewer experience.

It was easy to display suspicious transactions, but much harder to present them in a way that helps a reviewer understand what is happening quickly.

Another challenge was transforming raw transaction data into meaningful relationships. We wanted users to see how transactions connect through devices, IP addresses, merchants, and cards rather than viewing each alert in isolation.

(BACKEND CHALLENGES)************************************

We also had to prioritize features carefully due to the limited hackathon timeframe.

Accomplishments that we're proud of

Building a complete fraud review workflow
Creating explainable fraud alerts instead of black-box decisions
Designing the Fraud Constellation investigation network
Making fraud patterns easy to understand visually
Creating a polished and interactive user experience
Focusing on reviewer productivity rather than only detection accuracy

Most importantly, we built a platform that helps users understand fraud, not just find it.

What we learned

This project taught us that fraud detection is not only a technical problem but also a user experience problem. Obviously, we learned so much about data engineering in this challenge as we needed to clean, analyze, and detect data. We have never done this before so it was very interesting learning about new concepts.

We learned how important explainability is when presenting risk scores and suspicious activity.

We also learned the value of visualization. Showing relationships between transactions often reveals patterns that are difficult to notice in tables alone.

Finally, we learned how much can be accomplished when design, engineering, and product thinking work together.

What's next for 67Frauders

With more time, we would like to:

Add AI-powered investigation assistance
Generate automated investigation reports
Improve fraud pattern detection
Support real-time transaction monitoring
Add collaborative review workflows for teams
Expand the Fraud Constellation into a full investigation workspace
Introduce adaptive learning based on reviewer decisions

Our vision is to make 67Frauders a platform that helps trust and safety teams investigate fraud faster, understand risk more clearly, and make better decisions with confidence.

*VIDEO DEMO LINK: https://www.loom.com/share/2284ca933af646baad67f13cb274d009 *