From spreadsheets to reusable data
CBO publishes baseline projections for selected programs in Excel workbooks. Those files are valuable primary sources, but their layouts are designed for reading and can change between releases.
This project discovers and downloads those workbooks, inspects their structure, and transforms the data into versioned CSV datasets with schemas and release metadata.
Keeping the source attached
Each processed observation retains its source workbook, worksheet, row, and column. A verification stage reconciles processed values with the original files, and the pipeline stops when a stage fails.
The repository includes parsing instructions, generated catalogs, and tests so that updates can be repeated and inspected.