Python Data Pipeline Builder

Create a data processing pipeline for [data_description]. Requirements:
- Use pandas for data manipulation (or polars if already in project)
- Define clear pipeline stages: Extract → Validate → Transform → Load
- Add data validation at each stage with descriptive error messages
- Handle missing values: specify strategy (drop, fill, interpolate)
- Add type conversion with error handling (pd.to_numeric with errors='coerce')
- Include data quality checks (duplicates, outliers, schema validation)
- Add structured logging at each pipeline stage with row counts
- Implement retry logic for external data source connections
- Save intermediate results for debugging (optional, configurable)
- Write output to [destination: CSV/PostgreSQL/S3/BigQuery]
- Add CLI interface with argparse for running with different configs
- Include pipeline metrics: rows processed, duration, error count
- Type hints throughout, docstrings on public functions