BigCooking Data is a full-stack data science project. It features a food companion app that recommends a new recipe each and every day based on your taste. This application has been built from scratch without any structured data available for this purpose.

Visualization of the cluster of receipes created.
Visualizing clusters of receipes created using Principal Component Analysis.
  • Scraped over 100 000 cooking recipes from the web.
  • Pre-processed highly inconsistent data using NLP techniques.
  • Created new data by identifying recipe clusters using kmeans, PCA and NLP.
  • Developed a recommender system for recipes based on user-preferences and ingredient proximity.
  • Built a computer vision app that automatically register shopped food items.
Receipes recommended by our web-app.
Receipes recommended by our web-app.