feature-engineering-patterns

What problem does it solve?

This Skill provides expert patterns and best practices for feature engineering, helping you build more robust and performant machine learning models by avoiding common pitfalls and leveraging advanced techniques.

Core Features & Use Cases

• Feature Store Integration: Implement patterns for setting up and querying feature stores like Feast for point-in-time correct data.
• Robust Pipelines: Construct sklearn pipelines that correctly handle preprocessing and prevent data leakage.
• Time-Series Features: Generate lag, rolling, and cyclical features for temporal data.
• Target Encoding: Safely perform target encoding using K-Fold to avoid leakage.
• Drift Detection: Implement Population Stability Index (PSI) to monitor feature drift between training and production data.
• Anti-Patterns: Learn common mistakes to avoid in feature engineering.

Quick Start

Use the feature-engineering-patterns skill to generate time-series features for a given DataFrame.