wat-to-eat
Recipe recommendations from Food.com tags, in a Streamlit app.
Overview
A Streamlit app that recommends recipes from the Food.com recipe dataset. Pick a recipe and a tag group, and it returns the recipes whose tags in that group are most similar.
How it works
- Splits each recipe’s tags into groups such as time, cuisine, dietary, ingredients, taste and mood, occasion and technique.
- Vectorizes each tag group with scikit-learn’s CountVectorizer.
- Ranks recipes by cosine similarity to the selected recipe within the chosen group.
- Caches the data and loads the first 10,000 rows, to keep the app fast.
Features
- Recommendations by tag group, for example cuisine or dietary tags.
- A daily diet plan (breakfast, lunch, snacks, dinner and desserts) from your preferred ingredients and a dietary goal such as weight gain or muscle building.