Fraud Detection

Transaction anomaly and fraud risk intelligence.

Recommended GitHub Repo Name

transaction-fraud-detection-intelligence

Domain

FinTech security and transaction monitoring

Objective

Detect suspicious transactions using supervised and anomaly detection models.

Business Value

Reduces financial leakage and improves fraud operations response speed.

Dataset

Verified: https://www.kaggle.com/datasets/mlg-ulb/creditcardfraud

Run Locally

conda run -n ml-env streamlit run app.py

Dashboard Features

• Executive overview and business framing
• Prediction Center (single input + batch CSV)
• Model ranking leaderboard
• Model registry (all discovered model artifacts)
• Charts gallery (auto-load from charts/)
• Notebook traceability panel

Notebook Path Setup

Inside notebooks:

from path_setup import BASE, RAW, PROCESSED, MODELS, CHARTS, load_raw_csv

Project path helpers:

• path_utils.py
• notebooks/path_setup.py

Deploy on Render

This repo includes render.yaml. Connect repo in Render and deploy.

Start command used:

streamlit run app.py --server.port $PORT --server.address 0.0.0.0

Git LFS

Large assets are tracked with Git LFS (models, charts, data artifacts).