FoodData Central

This section explores FoodData Central, describing its structure and how it applies to culinary coaching.

Datasets considered

In order to ground our culinary coaching in reliable, open evidence, we reviewed a set of openly available food nutrition datasets. The goal was to identify a source (or sources) that are comprehensive, well‑maintained, and practical to work with in data science workflows.

FoodData Central (USDA)

https://fdc.nal.usda.gov/
A comprehensive database maintained by the U.S. Department of Agriculture. It integrates multiple data types (Foundation Foods, Branded Foods, Survey data, and more). The API is well documented, and the dataset is actively maintained. It is widely used in research and industry.

Open Food Facts

https://world.openfoodfacts.org/data
A global, crowd‑sourced database of food products, including nutrition facts and ingredient lists. While broad in coverage, data quality can be uneven because it relies on volunteer contributions. Particularly useful for packaged goods but less so for core ingredients.

European Food Information Resource (EuroFIR)

https://www.eurofir.org/
A network collating national food composition datasets across Europe. Access is partly commercial and partly open, with varying formats and limited direct API support. Much of the richer data requires membership or fees.

FAO/INFOODS Food Composition Databases

http://www.fao.org/infoods/infoods/tables-and-databases/en/
Global and regional datasets maintained by the FAO. These are valuable reference points, especially for international comparisons, but are published in static table form with limited programmatic access. While programmatic access is limited, these datasets remain authoritative for global comparisons.

Rationale for shortlist

From this review, FoodData Central was selected as the single shortlist candidate because:

• It provides both depth (detailed nutrient profiles) and breadth (Foundation, Branded, and other data types).
• It is actively maintained by a trusted government body (USDA).
• It has a stable API that is straightforward to use in Python and Observable workflows.
• Licensing allows open use for research, education, and product development.

While Open Food Facts is valuable for packaged foods and FAO/EuroFIR provide important international context, FoodData Central best meets the needs of this project: reproducible, high‑quality nutrition data that can be programmatically accessed and reliably maintained.