data-transform

The Data Transformation (Universal) skill provides a robust framework for cleaning, reshaping, normalizing, and preprocessing datasets using standard Python libraries such as pandas, numpy, and scikit-learn. It addresses the common challenges faced in data preparation, including handling missing values, removing duplicates, filtering outliers, scaling numeric features, and converting data formats. By performing these transformations locally, the skill ensures compatibility with any large language model (LLM) provider, offering flexibility and control without relying on cloud-based services.

This skill offers a comprehensive set of features for managing and transforming data efficiently. Users can normalize and scale numeric data using multiple methods, including Z-score, Min-Max, and Robust scaling, as well as apply log transformations for skewed distributions. It supports reshaping data between wide and long formats, merging multiple datasets, and creating new features from existing ones. The skill also facilitates cleaning operations such as filling or removing missing values, eliminating duplicates, and removing outliers using the IQR method. With these capabilities, it enables precise and repeatable preprocessing workflows tailored to specific datasets.

Ideal use cases include preparing raw data for machine learning models, statistical analysis, or reporting. Analysts, data scientists, and developers who work with structured datasets can leverage this skill to streamline preprocessing, ensure data quality, and enhance feature engineering. Its compatibility with any LLM provider makes it suitable for hybrid workflows where AI models require clean and structured input data, allowing users to maintain consistency and accuracy across different platforms and environments.