Open Food Facts Product Database
- Provider
- Open Food Facts contributors
- License
- ODbL-1.0 (source)
- Size
- 5 files
- Format
- Parquet (in sampled files)
- Stored at
- Hugging Face (https://huggingface.co/datasets/openfoodfacts/product-database)
- Access
- Free and public. No account with the data provider is needed.
- Category
- Health & Biomedical
- Source
- https://world.openfoodfacts.org/data
About Open Food Facts Product Database
The full product database of Open Food Facts, a collaborative, non-profit database of food products from around the world built by volunteers who scan barcodes and photograph packaging. The Hugging Face release converts the official JSONL dump into Parquet with pyarrow and is refreshed from it. Each row is one product keyed by its barcode, with brand, product name in several languages, quantity, categories, labels (organic, fair trade and so on), countries of sale, ingredients text and parsed ingredients, allergens and traces, additives, nutrition facts per 100 g, Nutri-Score, NOVA processing group, Green-Score/Eco-Score environmental indicators and packaging data. The root of the repository (measured 2026-11-14) holds food.parquet (about 8.1 GB), beauty.parquet (Open Beauty Facts, about 68 MB) and petfood.parquet (Open Pet Food Facts, about 8 MB). Uses include nutrition epidemiology, studies of ultra-processed food and food environments, product classification, ingredient parsing and recommender research. Read with pandas, polars, DuckDB or Hugging Face datasets. The database is licensed under the Open Database License, its individual contents under the Database Contents License, and product images under CC BY-SA.
How to load Open Food Facts Product Database in Python
from datasets import load_dataset
ds = load_dataset("openfoodfacts/product-database")
print(ds)Data structure
File types in sampled files: Parquet
Files
Example file: beauty.parquet
Analyze this dataset without downloading it
On Scigantic, Open Food Facts Product Database opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.
This page is also available as Markdown and as JSON.
Related Health & Biomedical datasets
- ISIC 2016 Part 1: dermoscopy skin lesion segmentation, 1,279 images with expert lesion masks (ISBI 2016 melanoma challenge)
- PUMA: melanoma H&E images with nucleus instances, ten nucleus classes and six tissue classes
- REDASA COVID-19 Open Data
- CT-ORG: multi-organ segmentation in whole-body CT (liver, lungs, kidneys, bladder, bone, brain)
- IDC CC-Radiomics-Phantom-3: Phantom CT (95 phantom subjects)
- CDC COVID-19 Forecast Hub: Weekly Probabilistic Forecasts of COVID-19 Hospital Admissions and Emergency Department Visits by U.S. State (Hubverse)
- UnumLocalia: same-section H&E, Xenium 5,001-gene spatial transcriptomics and COMET protein imaging of liver and tonsil
- Malaria Atlas Project: Plasmodium falciparum Malaria Death Count, 5 km annual maps