Sakshamvijay-078/fraud_detection_pathway ? reverse-engineered prompt

Build me a fraud detection system that can watch transaction data in real time, flag suspicious activity, and show the results in a clean dashboard.

I want the whole flow working end to end, with a streaming pipeline that takes in transactions, runs fraud checks and machine learning scoring, then saves the alerts and explanations so they can be reviewed in the app. The dashboard should feel modern and polished, with a dark mode look, loading states, and live updates for analysts to review suspicious cases.

Please set it up so I can run the data pipeline, the API, and the frontend locally with simple commands, and make sure the app works cleanly with the sample setup in the repo. If any model files or local setup details are missing, leave clear placeholders and tell me what I need to add. Look up current docs online if you need to.