Skip to main content
Intelligence Hub

USDA ยท Food Safety and Inspection Service

Microbiological Sampling Intelligence

Establishment-specific product sampling, kept separate from FSIS import-refusal data and interpreted within program and method boundaries.

Not connected

Official archives denied access

The FY2026 raw poultry and RTE ZIP links are listed and dated 14 October 2026, but both returned HTTP 403 with HTML instead of ZIP bytes. No login, bypass, guessed schema, or synthetic records were used. Zero rows are stored.

Source verified

FSIS publishes annual files from FY2014, updates the current year quarterly, and warns that laboratory methods change over time.

Import boundary ready

Isolated staging rejects schema drift, duplicate sample/analyte pairs and count mismatch before atomic promotion.

No misleading rate

A positive/tested fraction will appear only within one program, analyte and comparable method period, using positive plus negative samples as its denominator.

What will be supported after a verified import

Filters: fiscal year, product, project, analyte, result, and geography only where official fields support them.

Counting: samples remain distinct from organism-result rows; multi-organism testing never multiplies sample totals.

Comparability: method-change periods remain separate instead of being silently joined into a trend.

Meaning: establishment testing is not universal representative surveillance, product-risk prevalence, or evidence of human illness.