Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/shriram-vibhute/movie_recommendation_system

Movie Recommendation System - Web Application using Machine Learning: The system utilizes pre-processed movie data and a cosine similarity matrix to suggest similar movies
https://github.com/shriram-vibhute/movie_recommendation_system

bag-of-words machine-learning nlp nltk python recommender-system streamlit vectorization

Last synced: 3 months ago
JSON representation

Movie Recommendation System - Web Application using Machine Learning: The system utilizes pre-processed movie data and a cosine similarity matrix to suggest similar movies

Awesome Lists containing this project

README 

        
# MOVIE RECOMMENDATION SYSTEM 🎥🍿🎞️



![image](https://github.com/user-attachments/assets/4f6620bd-12af-4704-9b72-312fe6ea104e)






## **NOTEBOOK CODE**

Performed complete analysis of both the datasets



### Overview



This project implements a movie recommendation system using a dataset containing movie details and credits. The recommendation system suggests movies similar to a given movie based on features such as genres, keywords, cast, and crew.



### Project Structure



1. **Data**:

   - **movies.csv**: Contains information about movies, including budget, genres, keywords, overview, and more.

   - **credits.csv**: Contains information about the cast and crew of the movies.



2. **Libraries**:

   - **pandas**: For data manipulation and analysis.

   - **ast**: To parse string representations of lists and dictionaries.

   - **nltk**: For text preprocessing (PorterStemmer for stemming).

   - **sklearn**: For text vectorization and calculating cosine similarity.

   - **pickle**: For serializing objects to files.



### Data Processing Steps



1. **Importing Libraries**:

   Import necessary libraries for data manipulation, text processing, and machine learning.



2. **Loading Data**:

   Load the datasets using pandas.



3. **Merging Datasets**:

   Merge the `movies.csv` and `credits.csv` datasets on the movie title.



4. **Feature Selection**:

   Select essential columns for the recommendation system, such as `movie_id`, `title`, `overview`, `genres`, `keywords`, `cast`, and `crew`.



5. **Handling Missing Values**:

   Drop rows with missing values in the `overview` column.



6. **Feature Transformation**:

   - Transform the `genres` and `keywords` columns from string representations to lists.

   - Extract the first three actors from the `cast` column.

   - Extract directors from the `crew` column.

   - Split the `overview` text into a list of words.

   - Combine `overview`, `genres`, `keywords`, `cast`, and `crew` into a single `tags` column.

   - Apply stemming to the `tags` column to normalize words.



7. **String Vectorization**:

   Use `CountVectorizer` from sklearn to convert the `tags` into a matrix of token counts.



8. **Calculating Similarity**:

   Compute cosine similarity between movie vectors to determine similarity scores.



9. **Recommendation Function**:

   Implement a function to recommend movies similar to a given movie title based on the computed similarities.



10. **Saving Model**:

    Serialize the data and similarity matrix to files using pickle for future use.



### Running the Recommendation System



To get movie recommendations, use the `recommendation` function and pass the title of the movie you want recommendations for. This function will print out the top 10 movies similar to the provided title.






## **WEBAPP CODE**

Web Application related code



## Overview



This project is a movie recommendation system built using Streamlit, which provides an interactive web interface for users to get movie recommendations based on a selected movie title. The system utilizes pre-processed movie data and a cosine similarity matrix to suggest similar movies and display their posters.



## Project Structure



1. **Data**:

   - **movies.pkl**: A serialized dictionary containing movie data.

   - **similarities_matrix.pkl**: A serialized matrix of cosine similarity scores.



## Features



1. **Fetch Movie Posters**:

   - Retrieves movie posters from TMDb API using the movie ID. It includes retry logic for handling request failures.



2. **Get Movie Recommendations**:

   - Suggests movies similar to the selected movie based on cosine similarity scores.

   - Fetches and displays movie posters along with movie titles.



3. **Interactive Web Interface**:

   - Built using Streamlit to allow users to select a movie and view recommendations.






### APPLICATION VIEW



![image](https://github.com/user-attachments/assets/3af1b78f-3c9c-4476-bfb0-f358209c5bfc)



## Installation



To run this application, you need to have Python installed along with the required libraries. Install the dependencies using:



### Notes



- The recommendation system uses cosine similarity to measure how similar movies are based on their features.

- Proper feature transformation and text preprocessing are crucial for generating accurate recommendations.