Data Extraction Task

We are building a panel dataset of U.S. state governor-proposed budgets covering all 50 states from FY1995 to FY2027. For each state-year, we need line-item General Fund (GF) revenue (by tax type) and expenditure (by department/agency) extracted from official governor's budget PDFs. We have mostly collected the PDFs and extracted aggregate totals (one number for total GF revenue, one for total GF expenditure) for most state-years. This task is the next step: breaking those totals into their component line items -- individual tax categories for revenue, and individual departments or agencies for expenditure. For each state-year PDF, you will: 1. Open the governor's proposed budget PDF 2. Find the GF revenue table (revenue by tax type / source) and the GF expenditure table (appropriations or spending by department / agency / function). 3. Transcribe every line item exactly as printed, including subtotals and grand totals. 4. Record each value in millions (converting from dollars or thousands if needed). 5. Enter the data into a CSV with standardized columns (template provided). Qualifications 1. Careful attention to detail -- this is transcription work where accuracy matters 2. Comfortable reading government budget documents (PDFs with tables) 3. Comfortable with spreadsheets / CSV files Ability to work independently and flag ambiguities rather than guessing Note: Revenue and expenditure are sometimes on the same table (e.g., CO's "General Fund Overview" has revenue on top and expenditure below). Other times they are in completely different sections of the document (e.g., IL has revenue in Part II and expenditure in Part I, 40+ pages apart).