https://github.com/cintia0528/data_science-recommender_systems
Developing recommendation systems to replicate NETFLIX's user experience. Implemented Popularity Ranking, Memory-Based Collaborative Filtering (User-Based and Item-Based), and a Random Recommender.
https://github.com/cintia0528/data_science-recommender_systems
recommendation-algorithms recommendation-engine recommender-system
Last synced: about 1 year ago
JSON representation
Developing recommendation systems to replicate NETFLIX's user experience. Implemented Popularity Ranking, Memory-Based Collaborative Filtering (User-Based and Item-Based), and a Random Recommender.
- Host: GitHub
- URL: https://github.com/cintia0528/data_science-recommender_systems
- Owner: Cintia0528
- Created: 2023-12-14T12:02:17.000Z (over 2 years ago)
- Default Branch: main
- Last Pushed: 2023-12-14T13:31:28.000Z (over 2 years ago)
- Last Synced: 2025-03-31T05:35:16.858Z (about 1 year ago)
- Topics: recommendation-algorithms, recommendation-engine, recommender-system
- Language: Jupyter Notebook
- Homepage:
- Size: 1.85 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# WBSFLIX Recommendation Systems
## Goal
The goal of this project is to **develop recommendation systems** for Ursula, the owner of a hypothetical company named WBSFLIX, aiming to go digital. The aim is to replicate the user experience of popular platforms like NETFLIX. **Four distinct recommendation systems were implemented**: Popularity Ranking, Memory-Based Collaborative Filtering (User-Based and Item-Based), and a Random Recommender.
## Overview
WBSFLIX Recommendation Systems provide users with personalized movie suggestions based on different algorithms.
1. The Popularity Ranking system suggests trending movies,
2. User-Based Recommender recommends movies similar to a selected one based on user preferences using cosine similarity
3. Item-Based Recommender suggests movies based on the similarity between items, providing alternatives similar to the chosen movie
4. Random Recommender adds an element of surprise to the recommendations.
## Approach
1. **Popularity Ranking:** Identified trending movies based on a combined metric of average ratings and the number of ratings.
2. **Memory-Based Collaborative Filtering (User-Based):** Recommended movies similar to a selected one, considering user preferences and using cosine similarity.
3. **Memory-Based Collaborative Filtering (Item-Based):** Suggested movies based on the similarity between items, providing users with alternatives similar to their chosen movie.
4. **Random Recommender:** Introduced an adventurous element by randomly recommending movies, diversifying the user experience.
## Deliverables
The project includes a **Google Colab notebook** [here](https://github.com/Cintia0528/Project-9-Recommender-Systems/blob/3d5e876b1598180a1885a3c049a9061fcac8bf32/Movie_Recommenders_Final.ipynb) containing the code for all recommendation systems, a **Python notebook** [here](https://github.com/Cintia0528/Project-9-Recommender-Systems/blob/3d5e876b1598180a1885a3c049a9061fcac8bf32/app.py) including the Streamlit code and the [URL]() for the deployed website.
Since Streamlit glitches quite often, I also included a video of the finished product.
## Skills and Tools
- Python
- Streamlit
- Pandas
- NumPy
- Surprise
- scikit-learn
## Further Analysis
To enhance recommendation accuracy, in the future I aim to consider the following:
- **Enhanced Collaborative Filtering:** Explore advanced collaborative filtering techniques such as matrix factorization methods.
- **Content-Based Filtering:** Integrate content-based filtering to incorporate movie characteristics into recommendations.
- **Hybrid Models:** Combine multiple recommendation algorithms to leverage their strengths and improve overall accuracy.