Methodology
Data Source
All data on OpenSubsidies comes from the USDA Farm Service Agency (FSA) payment files, publicly available through the FSA Electronic Reading Room. These files are published as Excel (.xlsx) flat files containing individual payment records.
Data Processing Pipeline
- Download: We download all available Excel payment files from the FSA website. Our current dataset includes 83 files covering 2017–2025.
- Parse: Each Excel file is parsed using the
xlsxnpm library. Each row represents a single payment with 16 fields including recipient name, address, state, county, amount, program code, and program year. - Clean: State abbreviations are mapped to full state names. County names are standardized. Program codes are matched to program descriptions.
- Aggregate: Payments are aggregated by state, county, program, and recipient. Top recipients are identified by summing all payments to the same name within the same state.
- Output: Structured JSON files are generated for the web application — one for each dimension (states, counties, programs, recipients, yearly trends).
Key Definitions
- Payment: A single disbursement from the USDA to a recipient for a specific program.
- Recipient: The entity receiving the payment — may be an individual, partnership, corporation, LLC, trust, or estate.
- Program: The USDA accounting program under which the payment was made (e.g., CRP Annual Rental, ARC-County, Emergency Commodity Assistance).
- Program Year: The fiscal year attributed to the payment, which may differ from the disbursement date.
Limitations
- Recipient matching: Recipients are matched by name and state. The same entity may appear under slightly different names (e.g., "SMITH FARMS LLC" vs "SMITH FARMS, LLC"), which may lead to some undercounting of individual totals.
- Historical coverage: Our current dataset covers 2017–2025. The FSA publishes data going back to 1995, and we plan to backfill historical data over time.
- Negative payments: Some payments are negative (refunds, corrections, clawbacks). These are included in all totals.
- Crop insurance: Federal crop insurance subsidies (administered by RMA, not FSA) are NOT included in this dataset. They represent a separate, large category of farm support.
Update Schedule
FSA publishes new payment files periodically as disbursements are made. We check for updates monthly and reprocess the entire dataset when new files are available.
Questions?
For questions about our methodology or data, contact us at info@thedataproject.ai. See also our FAQ.