Film Recommendation System

This low-rank, collaborative movie recommendation engine yielded 10 movie recommendations of which 70% were consistent with user preferences versus a hierarchical clustering approach that produced only 3 recommendations out of 10 that were deemed watchable. The low-rank approach relies on the family of matrix decomposition/factorization techniques popularized during the Netflix recommendation engine challenge of 2006. Technology used included Python, SciPy, NumPy, Pandas, Seaborn, Matplotlib, Jupyter.