This project analyzes the interactions between users and articles on the IBM Watson Studio platform and makes recommendations to users about new articles. Three recommendation methods were explored: Rank Based Recommendation, User-User Based Collaborative Filtering and Matrix Factorization using Singular Value Decomposition (SVD).
Recommendations_with_IBM.ipynb
: Jupyter notebook containing main implementation and analysis.
data/user-item-interactions.csv
: Csv file with user-item interactions.
data/articles_community.csv
: Csv file with indexed items.
The project is organized along the following steps:
- Exploratory Data Analysis
- Rank Based Recommendations
- User-User Based Collaborative Filtering
- Matrix Factorization using Singular Value Decomposition (SVD)
The data were provided by IBM as part of Udacity Data Science Nanodegree program.