evidently-drift-detector

# Evidently Drift Detector Detect data drift, monitor model performance, and generate automated reports using Evidently AI. ## Overview This skill provides comprehensive capabilities for ML monitoring using Evidently AI. It enables detection of data drift, concept drift, target drift, and model performance degradation in production ML systems. ## Capabilities ### Data Drift Detection - Feature-level drift detection - Dataset-level drift analysis - Multiple drift detection methods (KS, PSI, Wasserstein, etc.) - Distribution visualization - Drift magnitude quantification ### Model Performance Monitoring - Classification metrics tracking - Regression metrics tracking - Performance degradation detection - Slice-based analysis - Error analysis ### Target Drift Analysis - Target distribution changes - Label drift detection - Prediction drift monitoring - Class balance monitoring ### Automated Reporting - HTML report generation - JSON metrics export - Dashboard integration - Custom metric creation - Test suite execution ### Production Monitoring - Real-time monitoring integration - Alerting threshold configuration - Time-series drift tracking - Batch comparison analysis ## Prerequisites ### Installation ```bash pip install evidently>=0.4.0 ``` ### Optional Dependencies ```bash # For Spark support pip install evidently[spark] # For specific visualizations pip install plotly nbformat ``` ## Usage Patterns ### Basic Data Drift Report ```python from evidently import Colum...