From PDFs to Pipelines: Liberating Public Data

Too much valuable information is trapped in documents. Here is the pipeline we use to set it free:

1. Collect — gather source documents and record provenance
2. Extract — parse tables and text into raw records
3. Standardize — map to a shared schema
4. Publish — release versioned, machine-readable data

Each stage is boring on its own. Together, they are how a commons gets built.