How wkgit0/-pakistani-indian-dishes-recipe-recommendation-system-by-ingredients works
Project type and purpose
• This is a data-science–driven recommendation system (recipe recommender) implemented as a Jupyter Notebook.
• Purpose: take a list of ingredients from the user, compute TF-IDF vectors over recipe titles/ingredients/steps, apply cosine similarity, and return ranked Pakistani/Indian dishes.
File structure
• Top level:
– “Pakistani Dishes Recipe Recommendation System by Ingredients.ipynb” (core code)
– “README.md” (project description)
• README mentions a data folder (“data/Pakistani_Dishes_Enhanced.xlsx”) even though it’s not in the file tree—this is the primary dataset.
• No explicit web service or frontend directories exist; the project lives in a notebook environment.
README for additional insights
• Model: TF-IDF vectorization + cosine similarity
• Weighted similarity across multiple fields (title, ingredients, steps)
• Dependencies: pandas, scikit-learn, numpy, openpyxl