About
About
About Recipe Database
Recipe Database is a data-driven food project that analyzes 197 traditional recipes from 195 countries through the lens of modern nutritional science. Every recipe is scored, compared, and broken down to the micronutrient level using data from the USDA FoodData Central database.
What Makes Us Different
- Every recipe has a Health Score (0–100) — computed algorithmically from nutrient density, healthy fat ratio, protein quality, Mediterranean alignment, and penalty factors for excess sodium, saturated fat, and added sugar.
- Every recipe has a "Lighter" version — algorithmically optimized to improve the health score while preserving the dish's cultural identity. We show exactly what changed and why.
- 21 nutrients tracked per serving — from macros (protein, carbs, fat) to micronutrients (iron, calcium, folate, vitamin A/C/E, omega-3, potassium, magnesium).
- 8 clinical badges — deterministic pass/fail criteria for Low Sodium, Heart Healthy, Diabetic Friendly, High Protein, High Fiber, Under 500 kcal, Mediterranean Aligned, and Potassium Rich.
- Ingredient-level impact analysis — every ingredient is scored for its delta contribution to the overall health score.
Data Sources
- USDA FoodData Central — Foundation Foods, SR Legacy, and Survey Foods databases for nutrient composition.
- FDA Daily Values (2020) — reference intakes for % Daily Value calculations based on a 2,000 kcal diet.
- Traditional recipes — sourced from culinary references, food historians, and verified against ingredient availability in each country.
Technology
The site is built as a static data application with pre-computed nutritional analysis. Recipe images are generated using FLUX.1 schnell, a state-of-the-art image generation model, with culturally-specific prompts for each country's traditional serving ware and setting. The interactive globe uses Three.js for 3D rendering.