recipe visualization

[marsh plume]

Recipe Classification and Visualization API
Welcome to the Recipe Classification and Visualization API project! This project harnesses the power of OpenAI's GPT-4o-mini and embeddings to categorize and visualize recipes submitted by users in the OpenAI Discord Grand Games.

Project Overview
This API project reads recipe submissions from a text file, determines if each submission is a recipe, and categorizes it as vegan, vegetarian, or neither. Using OpenAI's embedding model, recipes are converted into high-dimensional vectors and then reduced to two dimensions using t-SNE for visualization. The categorized recipes are plotted on a 2D scatter plot with interactive hover features for detailed exploration.

Key Features
Recipe Categorization: Automatically identifies and classifies recipes as vegan, vegetarian, or neither.
Dimensionality Reduction: Utilizes t-SNE to convert high-dimensional embeddings to 2D.
Interactive Visualization: Displays categorized recipes on a 2D scatter plot with color coding and hover annotations for easy analysis.
How to Use
Prepare Your Data: Ensure your recipe submissions are in a text file, starting each with "Challenge Submission:".
Run the API: Process the recipes and generate embeddings.
Visualize: Explore the categorization through the interactive 2D scatter plot.
Share your culinary creativity and gain insights into your recipe submissions with this powerful and interactive tool!

[marsh plume]

To use either run app.py (you have to copy the entire submissions chat into a txt first) which require your OpenAI AP Key and might cost a few cent, or simply just download and open the already created recipe-visualization.html file

[livid iris]

Love it, thanks for sharing!

[frail flame]

I’ve been wondering: what are these 2D plots supposed to enable us to do? I see them often but I don’t understand the user story.