"Predict the popular year for a song."
In this project, we will attempt to discover the features behind the popular music of each generation. For instance, if Britney Spears, “Oops I Did It Again” made the charts in 2001, and The Beatles’ “Real Love” made the charts in 1996, we want to see what made the music popular back then – was it the timbre, audio quality, or lyrics? We will then attempt to build a model that is able to predict when the song would have been most popular, which can be useful for future music analyses.
- Obtain the Million Song Dataset, or Million Song Subset: http://labrosa.ee.columbia.edu/millionsong/pages/getting-dataset
- Obtain the sample codebase. https://github.com/tb2332/MSongsDB
- Download SQLite DB Browser: http://sqlitebrowser.org
- Download Panoply (to open HDFS5): http://www.giss.nasa.gov/tools/panoply/download_mac.html
- Clone the repo, and work in dataexploration.py. Set your own path to the Million Song Dataset subset in setup.py.
Either:
- Use setuptools:
easy_install -U pyechonest - Download the zip from the releases page: https://github.com/echonest/pyechonest/releases
- Unpack, navigate to the echonest directory and run:
sudo python setup.py install
- Unpack, navigate to the echonest directory and run:
Thierry Bertin-Mahieux, Daniel P.W. Ellis, Brian Whitman, and Paul Lamere. The Million Song Dataset. In Proceedings of the 12th International Society for Music Information Retrieval Conference (ISMIR 2011), 2011.