JC
Back to projects
•
1 min read

Music Recommender System

A PySpark recommender system for large-scale music listening data stored in HDFS.

This big data project uses PySpark to extract and transform large-scale music listening data stored in HDFS and build a collaborative-filtering recommender system.

  • Developed a recommender system using an Alternating Least Squares model.
  • Evaluated the model against a popularity baseline.
  • Used Mean Average Precision at K as the main performance metric.
  • Improved MAP@100 by 16.7x compared with the popularity baseline.