USDA ERS Food Availability (Per Capita) Data System

It is the federal government's consumption ledger, kept since 1909. Direct measurement of what a nation eats is impossible at scale, so the USDA Economic Research Service does the next best thing: it accounts for the food itself. Beginning stocks, plus production, plus imports, minus exports, shipments to US territories and ending stocks - whatever disappears must have been eaten, hence residual disappearance, divided by the July 1 population and expressed as pounds per person per year.

More than 200 commodities ride the same arithmetic - red meat (beef, veal, pork, lamb, mutton), poultry (chicken and turkey), fish and shellfish, dairy, eggs, fats and oils, fruit, vegetables, grains, peanuts and tree nuts, added sugars and sweeteners - and a loss-adjusted companion series pushes further, subtracting spoilage and plate waste to approximate actual intake. This depth is why the record anchors the Packaged Foods & Meats catalog. Get a sample of this dataset.

A complete supply balance, not just a headline number. Each commodity-year carries beginning stocks, imports, total supply, exports, shipments to US territories and ending stocks - so the availability figure can be audited, rebuilt or re-decomposed rather than trusted blind. Meat commodities add three conversions (carcass, retail, boneless) that translate hanging weight into shopping-cart terms, plus the population base that turns tons into pounds per person.

That structure is the product. When chicken availability jumps, the components tell you whether imports surged, stocks drew down or production expanded - one table answers questions that usually need three datasets. Get a sample of this dataset.

Two shapes ship under one name. The historical series runs long - one measure, one year, one value, six columns - while the red-meat workbook runs wide, one row per year with the whole supply balance laid out side by side. Both arrive as flat columns that drop into any pipeline without reshaping.

Thirteen documented columns per commodity-year: the clock, the six-component supply balance, the three carcass-to-retail conversions, and the payoff pair of per-capita availability and its population denominator. The loss-adjusted companion adds consumer-level edible weight and calorie estimates. Definitions below are verified against published documentation.

Geography - United States, national averages only. No state or regional breakouts exist anywhere in the system; local questions need a companion dataset.

Temporal - 1909 to the present, with 1909-1989 held in separate Historical tables. Fish and shellfish received recent revisions covering 1990-2022, and loss-adjusted tables reach back over recent decades.

Granularity - one observation per commodity per year, in pounds per person. More than 200 commodities share one schema, so relative-value work across beef, broilers and canned peaches needs no stitching.

API, files, or your warehouse. Daily, weekly, or hourly.

Channel and cadence are settings, not projects. Files suit the analyst loading the full century once and versioning it beside their own sales master. Structured feeds suit products resolving commodity lookups inside an app. Warehouse delivery suits teams running per-capita screens in SQL against their own demand tables. Every delivery ships the field dictionary above, the captured sample rows and a coverage profile mapped to the commodities and periods you name - so the schema in the sample is the schema you ship against.

1. Protein processors and feed makers read the century-long beef, pork and broiler lines for capacity planning and domestic-pull forecasts (investors quants use cases).
2. Category strategists settle growth-versus-saturation arguments with pounds-per-person trends rather than vendor deck charts (market researchers use cases).
3. Trade economists decompose any commodity-year into domestic supply versus import dependence straight from the balance components.
4. Nutrition researchers take loss-adjusted edible weights and calories as intake baselines for dietary-
5. trend literature.
6. Sustainability analysts derive the loss wedge - core minus loss-adjusted - as their food-waste input.
7. Economists pair these quantities with price series to split volume moves from inflation (data scientists use cases).

Market researchers and consultants lead, because official quantities de-risk every sizing deck in the category. Investors and quant researchers gain a century-long annual panel of the protein complex that no commercial source matches for depth. Data scientists and ML engineers join a flat year-by-commodity table to anything they already model in one statement. Journalists, academics and students inherit the citable spine of American diet history, and policy and sustainability analysts get loss-adjusted estimates sturdy enough to stand in public commentary.

Within Packaged Foods & Meats this record owns quantity over a century - nothing else pairs 200+ commodities with a 1909 starting line. Its nearest neighbor, the USDA ERS Meat Price Spreads, owns the price side of the identical pound: farm, wholesale and retail values month by month since 1970. Analysts run both - quantity for structural demand, prices for margin - and the head-to-head sets out the trade-off explicitly. For forward-looking views, the USDA ERS Food Price Outlook adds forecasts this historical record deliberately lacks. See the best Packaged Foods & Meats datasets ranking.