This project focuses on building a Netflix Recommendation Engine from scratch. The goal is to develop a system that predicts which movies a user is likely to enjoy based on their preferences and rating history. The project leverages collaborative filtering and exploratory data analysis to understand user behavior and recommend movies in various genres accordingly.
In the notebook, data preprocessing, visualization, and recommendation logic are implemented.
Rating Distribution: Histogram and stats to understand overall rating trends.
Genre Analysis: Breakdown of how popular each genre is.
User-Movie Matrix: Pivoted table for collaborative filtering.
Similarity Measures: Cosine similarity used to recommend movies based on user preferences.
Drama and Comedy genres are among the most frequently watched and rated.
Certain movies have significantly higher average ratings, indicating clear user preference trends.
Sparsity exists in the user-movie matrix, meaning not all users have rated all movies (common in recommendation problems).
Collaborative filtering shows promising recommendations for users with similar movie rating histories.
A basic Recommendation Engine has been successfully created using user ratings and genres.
The system is capable of recommending top-rated movies for users based on similar users' preferences using collaborative filtering.
This approach simulates real-world strategies used by OTT platforms like Netflix and Amazon Prime to increase user engagement.