Name: Global Food Nutrition Database(10k Products)
Creator: Kaggle
License: http://opendatacommons.org/licenses/dbcl/1.0/

About this Dataset

Global Food Nutrition Database(10k Products)

9,804 Global Products with Nutri-Score, Eco-Score, NOVA Groups, Complete Macronutrients & Ingredients

Dataset Overview

This dataset contains a curated, high-quality snapshot of 9,804 global food products, aggregated from the Open Food Facts database as of December 2025. It provides a comprehensive multi-dimensional view of food health through three validated scoring systems:

Nutri-Score (A-E): Official European nutritional quality rating
Eco-Score (A+ to F): Environmental sustainability assessment
NOVA Classification (1-4): Food processing intensity scale

Key Statistics

Total Rows: 9,804
Total Columns: 20
Nutri-Score Coverage: 99.97% (9,801 products)
Eco-Score Coverage: 99.98% (9,802 products)
NOVA Group Coverage: 91.38% (8,959 products)
Ingredients Text Coverage: 95.94% (9,406 products)
Core Macronutrients Coverage: 93-95% (energy, fat, carbs, protein, sugars)
Unique Brands: 3,501 (Nestlé, Danone, Heinz, Hacendado, Tesco, Carrefour, etc.)
Geographic Scope: Predominantly European (France 20%, UK 6.5%, Spain 1.9%)

Data Science Applications

Nutri-Score Prediction (Multi-Class Classification)
Predict nutritional quality grades (A-E) based on macronutrient composition (fat_100g, sugars_100g, saturated-fat_100g, salt_100g).
Ultra-Processed Food Detection (Binary Classification)
Identify NOVA Group 4 products (53.5% of dataset) using ingredient lists and nutritional patterns.
Ingredient NLP & Allergen Detection
Extract allergens, E-numbers, and food additives from ingredients_text using Named Entity Recognition.
Environmental Impact Clustering
Group products by ecoscore_grade and nova_group to identify sustainable food categories.
Calorie Regression
Predict energy-kcal_100g from macronutrient breakdown.
Brand Health Profiling
Rank 3,501 brands by average Nutri-Score and percentage of ultra-processed products.
Category-Based Benchmarking
Analyze nutritional performance across 7,408 unique food categories.

Column Descriptors

Identification & Metadata (5 columns)

Column	Type	Coverage	Description
`code`	int64	100.00%	Unique barcode/EAN identifier (e.g., 6111035000430). Primary key.
`product_name`	object	97.57%	Commercial product name as displayed on packaging.
`brands`	object	96.19%	Manufacturer/Brand name (e.g., Nestlé, Danone, Heinz, Hacendado, Tesco).
`countries`	object	99.98%	Countries where product is sold (may include multiple, e.g., "France, United Kingdom").
`quantity`	object	89.92%	Package size/weight (e.g., "500 ml", "100 g", "2 L"). Mixed formatting.

Category & Labels (2 columns)

Column	Type	Coverage	Description
`categories`	object	98.77%	Hierarchical product categories (comma-separated), e.g., "Beverages,Waters,Mineral waters". 7,408 unique combinations.
`labels`	object	77.03%	Certification/Quality labels (e.g., "Organic", "Fair Trade", "Green Dot"). Multiple labels per product. 5,687 unique combinations.

Health & Environmental Scores (3 columns)

Column	Type	Coverage	Description
`nutriscore_grade`	object	99.97%	Official Nutri-Score (A=Healthiest to E=Least Healthy). Includes "not-applicable" (2.98%) and "unknown" (8.20%). Distribution: A=16.11%, B=12.70%, C=22.48%, D=18.23%, E=19.28%.
`ecoscore_grade`	object	99.98%	Environmental impact score (A+=Best to F=Worst). Considers carbon footprint, water use, packaging recyclability. Distribution: A+=5.41%, A=15.46%, B=18.59%, C=13.39%, D=14.29%, E=10.18%, F=2.52%.
`nova_group`	float64	91.38%	NOVA processing classification (1=Unprocessed, 2=Processed ingredients, 3=Processed foods, 4=Ultra-processed). Distribution: Group 1=11.02%, Group 2=4.15%, Group 3=22.69%, Group 4=53.52%.

Ingredient Data (1 column)

Column	Type	Coverage	Description
`ingredients_text`	object	95.94%	Full ingredient list (unstructured text). Multi-language (English, French, Spanish, Arabic). Contains allergens, E-numbers, additives. Perfect for NLP/NER tasks.

Macronutrients per 100g (9 columns)

Column	Type	Coverage	Mean	Median	Description
`energy-kcal_100g`	float64	94.66%	292.75	280.0	Energy in kilocalories per 100g. Waters = 0 kcal, oils ≈900 kcal.
`fat_100g`	float64	94.95%	15.32	7.02	Total fat content per 100g. Low-fat: <3g, high-fat: >17.5g (EU thresholds).
`saturated-fat_100g`	float64	93.23%	5.25	1.50	Saturated fatty acids per 100g. Key penalty factor in Nutri-Score calculation.
`carbohydrates_100g`	float64	94.70%	30.19	18.0	Total carbohydrates per 100g (sugars + fiber + starch).
`sugars_100g`	float64	93.51%	11.98	4.20	Total sugars per 100g (glucose, fructose, lactose, sucrose). High-sugar threshold: >22.5g.
`fiber_100g`	float64	69.74%	4.09*	2.90	Dietary fiber per 100g. *Note: 2 outlier products exceed 100g/100g (data quality issue). Mean calculated after excluding outliers >50g.
`proteins_100g`	float64	94.90%	7.33	6.20	Protein content per 100g. High-protein products: >20g/100g.
`salt_100g`	float64	93.24%	1.24	0.39	Sodium chloride (salt) per 100g. WHO guideline: <5g/day total intake.
`sodium_100g`	float64	93.24%	0.50	0.16	Elemental sodium per 100g. Conversion: Salt (g) = Sodium (g) × 2.5.

Ethically Mined Data

Source: Open Food Facts (https://world.openfoodfacts.org)
License: ODbL (Open Database License) – Free to share and adapt with attribution
Collection Method: Community-contributed data + Official manufacturer information
API Retrieval: December 2025 snapshot via Open Food Facts API v2
Privacy: No personal data – Product-level information only
Transparency: All data traceable to Open Food Facts database with full revision history

Acknowledgements

Data Source: Open Food Facts collaborative database
Curator: Kanchana Karunarathna (Kanchana1990)
Community: 300,000+ Open Food Facts contributors worldwide
Scoring Systems: Santé Publique France (Nutri-Score), Open Food Facts Foundation (Eco-Score), NOVA Classification (University of São Paulo)

Tables

Openfoodfacts Nutrition Final 2025–12–10

@kaggle.kanchana1990_global_food_nutrition_database10k_products.openfoodfacts_nutrition_final_2025_12_10

2.51 MB
9,804 rows
20 columns

CREATE TABLE openfoodfacts_nutrition_final_2025_12_10 (
  "code" BIGINT,
  "product_name" VARCHAR,
  "brands" VARCHAR,
  "countries" VARCHAR,
  "quantity" VARCHAR,
  "categories" VARCHAR,
  "labels" VARCHAR,
  "nutriscore_grade" VARCHAR,
  "ecoscore_grade" VARCHAR,
  "nova_group" DOUBLE,
  "ingredients_text" VARCHAR,
  "energy_kcal_100g" DOUBLE,
  "fat_100g" DOUBLE,
  "saturated_fat_100g" DOUBLE,
  "carbohydrates_100g" DOUBLE,
  "sugars_100g" DOUBLE,
  "fiber_100g" DOUBLE,
  "proteins_100g" DOUBLE,
  "salt_100g" DOUBLE,
  "sodium_100g" DOUBLE
);