data-integrate

What problem does it solve?

This Skill resolves the common and risky task of joining multiple tabular sources without clear grain definition, preventing unintended duplicates, temporal mismatches, and undocumented transformations that break downstream analysis.

Core Features & Use Cases

• Grain enforcement: Forces explicit definition of the analysis unit (record grain) before any join to avoid semantic inconsistencies.
• Join planning & validation: Guides key null checks, cardinality inspection, and pre/post-join row-count comparisons to detect fan-out and other anomalies.
• Temporal alignment & leakage prevention: Validates time ranges and prevents future information leakage when merging time-series data.
• Integrity checks & lineage: Runs missing/duplicate/primary-key checks, records provenance into analysis_context.md, and documents data lineage for reproducibility.
• Persistence: Outputs a canonical analysis_table.csv and records execution logs and lineage into the project documentation for auditability.
• Use case: Combining customer, order, and event tables into a single analytics table while ensuring key uniqueness, time alignment, and traceable provenance.

Quick Start

Use data-integrate to merge your source tables into a single analysis_table.csv after confirming the analysis grain and recording lineage in analysis_context.md.