Inspiration
The inspiration for this project stemmed from a love for movies and the desire to help people discover new films based on their interests. With so many movies available, having a tool to recommend similar films based on one's favorites can enhance the movie-watching experience.
What it does
The project analyzes a dataset of movies and recommends similar films based on a selected movie. It uses natural language processing and machine learning techniques to compare movie features and suggest titles that closely match the input movie's characteristics.
How I built it
- Data Collection: Gathered movie data from CSV files.
- Data Preprocessing: Merged datasets and cleaned data using pandas.
- Feature Extraction: Extracted relevant features such as genres, keywords, cast, and crew, and combined them into tags.
- Text Processing: Applied stemming and vectorization using CountVectorizer to convert tags into a numerical format.
- Similarity Calculation: Used cosine similarity to measure the closeness between movie vectors.
- Recommendation System: Implemented a recommendation function that suggests similar movies based on the selected title.
- User Interaction: Added a fuzzy matching feature using the fuzzywuzzy library to handle movie title misspellings and provide user-friendly interactions.
Challenges I ran into
- Data Cleaning: Handling missing values and inconsistent data entries.
- Feature Extraction: Effectively extracting and combining different movie features into a meaningful representation.
- Text Processing: Ensuring accurate text preprocessing and stemming to improve the quality of feature vectors.
- User Interaction: Implementing fuzzy matching and handling user inputs in a way that enhances the recommendation experience.
Accomplishments that I'm proud of
- Successfully creating a comprehensive dataset by merging and cleaning multiple data sources.
- Developing an efficient recommendation algorithm that provides accurate and relevant movie suggestions.
- Implementing user-friendly features such as fuzzy matching and interactive prompts to improve user experience.
What I learned
- Advanced data preprocessing techniques using pandas and NumPy.
- Natural language processing methods for text feature extraction and stemming.
- Machine learning concepts related to vectorization and similarity measurement.
- Effective user interaction handling to enhance the usability of the recommendation system.
What's next for the Movie Recommendation System
- Real-Time Updates: Incorporating real-time movie data to keep the recommendations up-to-date.
- Enhanced Algorithms: Exploring more advanced algorithms like collaborative filtering and deep learning for improved recommendations.
- User Profiles: Allowing users to create profiles and get personalized recommendations based on their viewing history and preferences.
- Web Application: Developing a web-based interface to make the recommendation system more accessible.
- Multilingual Support: Adding support for multiple languages to cater to a wider audience.
Built With
- jupyter
- nltk
- numpy
- pandas
- python
- scikit-learn
- seaborn
- vs-code
Log in or sign up for Devpost to join the conversation.