About

This project is an autonomous analytical system that detects high-risk events in NASDAQ-100 stock prices. It uses an Isolation Forest model on engineered features (daily returns, 7-day rolling mean, and 7-day rolling standard deviation) to identify anomalies in historical stock data.The workflow is scheduled to run automatically daily, generating risk alerts labeled as “HIGH” or “LOW” based on detected anomalies. Validation is performed via historical backtesting, monitoring anomaly frequency (~3%), and visual inspection of detected anomalies against known market behavior.
This system demonstrates predictive analytics in production, delivering real-time insights with automated notifications, fully deployed using Zerve’s scheduling features.
The system runs daily at 11:00 IST, automatically detects anomalies in stock data, classifies risk levels, and produces alerts. Email notifications are enabled to inform stakeholders if the job succeeds or fails.