pranav14-1/cyber_threat_detection_pipeline ? reverse-engineered prompt

Build me this cyber threat detection app from the README. I want a Python based pipeline that takes in enterprise access logs, learns normal behavior, flags suspicious activity, and then explains why something looks risky in plain English.

It should combine a statistical baseline for unusual login patterns, a sequence model for weird event order, and a threat classifier that can label attacks like brute force, credential stuffing, impossible travel, low and slow exfiltration, and lateral movement. When an alert is raised, I want a clear explanation card that tells a SOC analyst what features drove the decision, plus a dark themed Streamlit dashboard where they can review ranked alerts, see location movement on a map, and leave feedback.

Use the existing project structure if helpful, make it runnable, and wire up whatever config, model files, and database schema are needed. If you need current docs for any libraries, look them up online.