Fin-Analysis

Fin-Analysis analyzes financial data to detect anomalies using forensic methods like Benford’s Law and AI-assisted analysis, delivering clear and explainable risk insights.

Comment

Fin-Analysis — My Project Journey

What Inspired Me

The idea for Fin-Analysis came from observing how financial fraud detection and data verification are still largely manual, complex, or locked behind expensive enterprise tools.
As a student working with data and analytics, I noticed a gap: there was no simple, beginner-friendly platform that could analyze financial data using real forensic techniques.

I wanted to build something that:

  • Uses actual auditing logic
  • Works with plain CSV / Excel files
  • Is understandable even for non-accountants

That curiosity led me to explore Benford’s Law and how it’s used in fraud detection — and that became the foundation of Fin-Analysis.

How I Built Fin-Analysis

The system is designed as a clean, modular, full-stack web application.

Frontend

  • Built using React + Vite
  • Focused on a clean, modern UI
  • Pages include:
    • Login & Signup
    • File Upload
    • Dashboard
    • Analysis History
    • Settings
  • Designed with usability in mind using clear KPIs, visual charts, and PASS/SUSPICIOUS indicators

Backend

  • Built using FastAPI
  • Handles:
    • File processing
    • Data parsing
    • Statistical analysis
    • Generating results

Analysis Engine

  • Core logic is based on Benford’s Law
  • For a dataset, the leading digit distribution is calculated and compared with the expected Benford distribution:

[ P(d) = \log_{10}(1 + \frac{1}{d}), \quad d = 1,2,\dots,9 ]

  • A deviation score helps classify data as PASS or SUSPICIOUS
  • Additional statistical summaries enhance insight quality

AI Integration

  • Gemini API is used to convert raw statistical outputs into human-readable insights
  • This makes the results understandable, not just technically correct

Cloud & Google Technologies

  • Firebase Authentication for secure login
  • Firebase Storage for file uploads
  • Firestore for storing analysis results and history
  • Google’s ecosystem ensures scalability and security

What I Learned

This project taught me more than just coding:

  • How forensic accounting concepts work in real life
  • Practical application of Benford’s Law
  • Designing user-focused dashboards
  • Handling file uploads, asynchronous processing, and APIs
  • Debugging real-world issues such as CORS, authentication errors, and cloud permissions
  • Structuring a project from idea to PRD, MVP, and final product

Most importantly, I learned how to translate complex math and data science into usable software.

Challenges I Faced

Some of the key challenges included:

  • Correctly configuring Firebase Authentication and Storage
  • Debugging API key and CORS-related issues
  • Ensuring uploaded files were processed reliably
  • Handling inconsistencies in CSV file formats
  • Balancing technical depth with simplicity for users
  • Skipping planned PRD features that were not MVP-ready while keeping the project coherent

Each challenge helped me understand the tools and systems more deeply.

Final Thoughts

Fin-Analysis is not just a project — it is an attempt to make financial integrity tools accessible.
It combines statistics, AI, cloud computing, and clean UI design into a real-world solution.

This journey reinforced one belief for me:

Good technology is not about complexity — it is about clarity.

Built With

Share this project:

Updates