Paul Lamere is preparing is own tutorial on map/reduce and the MSD. And since it's Paul, he blogs about every step of the way.
Paul Lamere (big shot at The Echo Nest, co-creator of the MSD) is preparing a SXSW panel (still a proposal?) on large-scale music data processing. What it means in practice? "What can you do with map/reduce and the Million Song Dataset?".
Paul is a great demonstrator, and his blog http://musicmachinery.com/ is quite famous. If you're not following it, we still recommend today's post:
How to process a million songs in 20 minutes
Paul gives you access to the s3 bucket with the million songs converted to text files, he gives you his code, he explains the map/reduce pricing on Amazon, he performs sanity checks on the result through 30-second clips... pretty complete demo! And there should be more to come!
Obviously we'll pay close attention to these posts and probably integrate some of them in our ISMIR tutorial in October. But we encourage you to give feedback: what would you like to see computed on the MSD? do you think you can make things even faster? If you were looking for an easy way to start using the MSD, this might be it!
Finally and unrelated, Montreal Music Hack Day is coming! and yes, Montreal is the greatest city on Earth.
Happy coding!
-TBM
- millionsong's blog
- Login to post comments