This is our project which we finished in 24 hours for Dragon Hackathon in University of Ljubljana.
This project aims to simplify the process of choosing movies for viewers. Making a decision on which movie to watch can often be time-consuming, so our project streamlines this experience. The main screen offers two options: the first provides 10 random movie recommendations, while the second allows users to specify actor, category, or language preferences (up to 3 selections in any combination) to receive tailored recommendations. These recommendations are generated using an adjusted pagerank algorithm applied to an affiliation network.
Upon receiving recommendations, users can click on a movie title to open a new page featuring a chatbot. Initially, the chatbot provides general information about the selected movie. Users can continue interacting with the chatbot to ask additional questions about the movie.
This project's flexibility extends beyond movie recommendations; it can be adapted to offer recommendations in any domain where affiliation network data is available.
Project is created with Python. The main packages used:
- Langchain (LLM)
- Numpy (Data analysis and cleaning)
- Crewai (Crawling Internet for Chatbot)
- Networkx (Network analysis)
- Streamlit (Web application)
