The purpose of this project is to create an entry to submit to a competition on Kaggle involving music recommendations. The goal is to use machine learning to generate predictions of whether a user will repeatedly listen to a song within a month of first listening to it. The data provided by Kaggle for this competition come from a music streaming service from Asia called KKBOX. To increase the accuracy of our predictions, we also plan to pull in data from other online streaming sources.
- Lauren Gripenstraw
- Angela Ho
- Qilin Liu
- Katie Chong
- Initial project planning
- Begin gathering data
- Finish gathering data
- Start cleaning data
- Finish cleaning data
- Begin employing machine learning techniques
- Continue employing machine learning techniques
- Continue with machine learning
- Begin to test using official test set provided by Kaggle
- Finalize predictions for official test set
- Prepare project for submission to Kaggle
- Submit to Kaggle
- KKBOX data provided by Kaggle
- Spotify API
- Dataset gathered from Last.fm API
- Pandora API
- Python for web scraping
- Python for machine learning *scikit-learn
Create a more accurate music recommendation engine with machine learning