Flavor Map
Recipe data → interactive map
Python cleans ingredients and scores their relationships across 231,000 recipes. JavaScript turns that data into an interactive map; Node.js prepares layouts and puzzles.
Python standardizes messy recipe data.
Co-occurring ingredients become weighted relationships.
JavaScript lets you follow pairings to real recipes.
Python · JavaScript · Node.js · Cloudflare Workers · Claude Code
How it works
A data pipeline cleans ingredient names from Food.com recipes and counts which foods appear together. A statistical association score ranks pairings. The site turns those relationships into a map and food-linking games, with recipe links behind the pairings.
What I used
Python for recipe preparation and pairing data, JavaScript for the interactive map and games, Node.js for layout and puzzle generation, and Cloudflare Workers for hosting. Claude Code helped build and refine the tools.
Why I built it
I wanted to explore food pairings without assuming that two foods go together just because both pair with a third ingredient.