How suyashagnihotri/bollywood-genre-based-movie-recommender works
Project Type and Purpose
• This is a standalone, desktop-based recommendation application built in Python.
• Purpose: Let users pick a Bollywood movie title from a GUI and get top-N genre-based suggestions computed via TF–IDF and cosine similarity.
Key Components
a. User Interface Layer
• CustomTkinter GUI (movie_gui.py)
• Widgets: dropdown/listbox for movie selection, “Recommend” button, results display area
b. Data Layer
• CSV data source (movies_metadata.csv)
• Data ingestion via Pandas
c. Processing / Algorithm Layer
• Text preprocessing (TF–IDF vectorization on the “genres” field)
• Similarity computation (cosine similarity using Scikit-learn)
• Recommendation logic (sorting and selecting top-N results)
d. Integration and Flow
• GUI triggers data load & processing
• Algorithm returns a ranked list of movie titles
• GUI presents the recommendations
Architectural Patterns & Design Principles
• Simple MVC‐style separation:
• Model: movies_metadata.csv + Pandas DataFrame
• View: CustomTkinter components
• Controller: recommendation logic and event handlers in movie_gui.py
• Single‐process, in-memory data pipeline—no external services.
• Lazy loading: load & vectorize dataset only once at startup for responsiveness.
Technology Stack Labels
• Python 3.x
• Pandas (data ingestion & manipulation)
• Scikit-learn (TF–IDF, cosine similarity)
• CustomTkinter (desktop GUI)
• CSV file as data store