Here you'll find all metadata collected and created during the project. For each of the datasets we created an RDF repository which you can download.
Jazz Encycplopedia (JE) is a collection of 500 CDs covering jazz history from its beginnings to 1950s. It is the cleanest dataset with metadata including tune primary and secondary titles with composers/arrangers as well as detailed session/track lineups. Lord is a larger collection of metadata downloaded from the Lord online database for the purposes of the project; yet the metadata are less reliable or sometimes missing. The Illinois dataset (ILL) is a corpus of CDs from Iniversity Urbana-Champaigne Jazz library, with associated Lord metadata, starting from 1960s. The 100 years of Jazz dataset is a union of JE and ILL with tunes, musicians, bands and instruments merged. DTL1000 is a clean subset of '100 Years of Jazz' where all tracks that have solos have solos segments associated with them.
The current versions of the RDF (ttl) files you'll find here, in the corresponding components.
The code that creates the datasets as well as the newest versions of original datafiles can be found on [GitHub][1]
Semantic model component contains the diagrams presented in the Jazz Ontology paper.
For more details see our papers.
[1]: https://github.com/ppquadrat/DigThatLick