There are several ways to access the data we produce:

SPARQL Endpoint

We provide a SPARQL endpoint (using a Squebi interface) through which to access the relationship and authority data we store. This includes all of the relationship triples and basic biographical information about the musicians we’ve found in the transcripts that have been processed.

For example, to find everyone that Mary Lou Williams knows of:

prefix lj: <>
prefix rel: <>

lj:Mary_Lou_Williams rel:knowsOf ?object
} LIMIT 20

To search for Mary Lou Williams’ URI:

prefix foaf: <>

?subject foaf:name ?object

FILTER regex(?object, “mary lou”, “i”)

Published URIs

As specified, all names discovered in transcript processing (and soon, via dataset mashups and processing of other documents besides transcripts) are given URIs in the Linked Jazz Namespace. The URIs consist of a URL address–in other words, they live (in LOD jargon, are “dereferenced”) at a specific location on the web. Human-readable versions of our URIs display in the URL bar with the word “page.” For example:


Machines can also read our URIs, but the address they read contains the word “resource.” For example:


We are using LodView to display our human-readable dereferenced URIs. Developed by the LodLive team, LodView is a java web application that draws from the triples available via our SPARQL Endpoint. It displays attributes of the resource (such as type) and name. It also shows inverse relations (in other words, relationships for which the URI is the object, not the subject of the triple). For more about LodView, see developer Diego Camarda’s github repository.

Machines can request Linked Jazz data in a few standard formats (N-triples, Turtle, XML, and JSON-LD). If you want to try, open up a Terminal and type one of the following:
curl -L -H “Accept: text/rdf+n3″
curl -L -H “Accept: text/turtle”
curl -L -H “Accept: application/rdf+xml”
curl -L -H “Accept: application/rdf+json”

Data Dumps

Linked Jazz provides dumps of people and relationship data in various formats. Relationship information has two facets:

  1. That a basic rel:knowsOf relationship exists and its direction.
  2. Annotations from the crowdsourcing tool as to the more specific nature of the relationship.

All annotations are returned for a specific relationship, these annotations are added to the system by visitors to the Linked Jazz 52nd St. crowdsourcing tool. Currently, we do not aggregate the varying opinions of a specific relationship into a quantified value. Rather, all opinions (even duplicate) are included in the dump. Data is available in triple format or as a Gephi file. Annotated relationships are not returned in the Gephi network, only a single basic rel:knowsOf. Data dumps are updated hourly.

Get all people in triple format:

Get all relationships in triple format (includes crowdsourcing data):

Get all relationships in a Gephi file (only includes rel:knowsOf relationships):



Our legacy API is available here.