The project consisted of a Dockerized, CLI-driven data pipeline for reconciling multi-source marketing data into a single canonical dataset. The system processed CSV inputs through stages: normalization, validation, reconciliation, conflict resolution, and reporting. I implemented matching logic using exact identifiers, mapping tables, and conservative heuristics; it used explicit metric precedence rules for resolving conflicts. The pipeline produced an HTML summary report along with audit artifacts.
Built with
- Python
- Data Engineering
- Docker