Music Park

Springsteen Number

Find the shortest collaboration path between two artists, or measure an artist's Springsteen Number.

Try one of these artists

Choose or type a source artist.

Shortest-path Graphs?

These are some of the additional paths that exist between the source and target artists. For example of the collaboration number is 4, there may be more than the one path that exists between the two artists. The graph shows some more options (but not necessarily all) of the shortest paths that exist between the two artists. This is whats called a Directed Acyclic Graph (DAG). Directed because the edges have a direction, and acyclic because there are no cycles (loops) in the graph. This is the graph algorithm being solved when finding multiple of the shortest paths between the nodes. Of course, this graph is actually bidirectional because collaborations can be viewed from either artist's perspective, but yeah, kind of a neat thing :)

What am I looking at?

Music Collaboration Number? Springsteen Number? What?

The Music Collaboration Number (MCN) measures the collaboration distance between two artists, based on shared recordings or band memberships. Think of it as a "degrees of separation" for musicians. Two artists who have directly collaborated have a distance of 1.

The Springsteen Number is a specific case of the MCN, measuring the collaboration distance from any given artist to Bruce Springsteen. This is similar to the concept of an Erdos number in mathematics, which measures the collaborative distance between mathematicians based on co-authorship of academic papers. Read more about these numbers in the README of the github repository linked below.

Server RepoSix Degrees of SeparationErdos-Bacon Numbers

How are you doing this?

TLDR: BFS on a Bipartite Graph. You can read more on the GitHub. The data for the 40M recordings, ~1.3M artists and ~840k relationships and more is built from MusicBrainz database dumps, and queried on my API.

MusicBrainz is an open collaborative music encyclopedia. This site attributes artist and recording metadata to MusicBrainz, and MusicBrainz also offers an API for looking up public data directly. They do this for free, so if you like this site or see any errors, consider supporting MusicBrainz by donating or contributing to their project :)

MusicBrainzMusicBrainz API

Why do you say recordings? Also some of them are wrong...

All of this data is from MusicBrainz, which goes beyond just songs/tracks you might find on Spotify. This includes all sorts of recordings, like live performances, bootlegs, etc. In rare cases, some of the data is subject to human error and therefore incorrect (and you can help fix it!).

Other times, attributions can be interesting. For example, O Little Town of Bethlehem (feat. Elvis Presley) by Pentatonix is a recording that features Elvis Presley, even though he passed away decades before the recording was made. Legally speaking, Elvis is a collaborator on this recording (whoever owns the rights to his voice made this decision), so it counts as a collaboration in the graph. This is also why you might see some unexpected collaborations in the results.

Another quirk of the graph algorithm is the way memberships are dealt with. Membership to a group does not bypass the group itself. For example, on Under Pressure, David Bowie collaborated with Queen. The distance from Bowie to Queen is 1, but the distance from Bowie to Freddie Mercury is 2, because the path goes through Queen. This takes care of some edge cases like bands that have completely different lineups over time.

Why Springsteen?

I'll talk about this soon :)