The first emotion-sensitive music API
to deliver top-tier intelligent music experience


Take profit of a music database owning the largest granularity
and emotional content to date.

Retrieve information, get recommendation, achieve psych-emotional profiling and matching

Musimap mapped the global music catalogue
as a neural music network encompassing:

  • 3 billion Data Points
  • 2 billion Relations
  • 40 million Songs with pre-analyzed moods
  • 4,5 million Artists
  • 1,6 million Songs manually-tagged
  • 75 thousand Artists manually-tagged
  • 100 weighted parameters per song on average
  • 11 thousand Keywords
  • 400 Genres
  • 1,5 thousand sub-genres
  • 300 Moods
  • 100 Listening Contexts
feature1

300 Moods

Musimap defined 6 main families of moods, divided each in 3 sub-families, which in turn are separated into 10 to 30 detailed moods, thus reaching a total of 300 nuanced moods. An approximate 3000 aliases (i.e. semantic words) are then associated to the nuanced moods, this way giving shape to the whole palette of human emotions. On average, a track is appointed 50 moods.

400 Genres

Genres are lexicological terms used to describe a united movement in the history of music.

Each music genre is part of a system of crossed-influences with all other genres. Each artist, album and track is also enclosed in such a network. To be qualified, each album is defined by a main genre and several influences from other genres. Each genre is also portrayed with keywords, moods, rhythmic moods, etc. in relation with its key artists (pioneers and ambassadors hereunder explained).

E.g. the genres Alternative rock, Glam Rock or Punk all have a weighted relation to Rock
E.g. an album by The Rolling Stones is identified in the main genre Rock and has influences of other genres such as Blues Rock, Rock ‘n’ Roll, Rhythm & Blues, Pop, etc.

40 million Tracks

A track is a composition included in an album or released as a single; it has a cover artwork and is performed by one or more artists.

Key information related to a track are the release date of the medium, its recording dates, recording locations, labels and publishers. A track is referenced with genres and the date (first recording or alternatively the date of release). A track has further information attached: keywords, moods, rhythmic moods, BPM, main key, chromatic range, situations, rhythmic signature (4/4, 3/4, etc.), instrumentation and performers.

A Track-Link is a link between two tracks where one of the track uses part of the other (as Cover, Remix, Medley, Sample, etc.).

Every released album and its tracks are distributed by music publishers and labels. A label is portrayed in relation to all its releases and by all the genres, artists, moods, keywords etc. it encloses.

4 million Albums

An album is a group of tracks whose author rights belong to one or different artists (referred to as ‘Owners’).

Key information related to an album are release dates, recording dates, recording locations, labels and publishers. An album is categorized in a main genre at a specific date (first recording or alternatively the release date). Albums are also referenced with keywords and genres, derived from the keywords and genres attached to all their tracks. References or links to the main music or data providers (e.g., Qobuz, Deezer, YouTube, Amazon, Wikipedia, etc.) are also indicated.

4.5 million Artists

An artist is a band or an individual associated to an album or a track either as performer or producer, composer, writer, technician, etc. Such artist can own several albums and perform directly on several tracks even if he / she does not own the album.

Each artist has its own profile defined by moods, keywords, languages, instruments, tags, etc. and belongs to our large network of crossed-influences between Genres, Artists, Albums and Tracks.

2 billion Influences

All types of relations between two items: e.g., artist to artist, genres to genres, genres to artist, artist to track, track to track, artist to genres, genres to track. Please note that these influences level is always indicated with a weighted percentage (%) and a date, as well as with direction: (influenced) From … / (influence) To …

E.g. FROM: influences the artist received from his / her cultural and educational background (it could be Styles, others artists, tracks, etc.).
E.g. TO: influences the track has on other styles and on other artists, albums, tracks, etc.

So, how does it work?

To use our recommender system, you can simply call our API (with ISRC/UID references or our references) or Musimap can match the tracks of your database (XML) while generating new data for any unmatched title via our integrated technology. Musimap’s recommender system can then be applied either to your music library specifically or to other specific portions of the global music catalogue.

By using Musimap’s API, you can:

  • retrieve exhaustive information on artists, albums and tracks (e.g., moods, voice families, etc.)
  • get recommendation (curation at scale, evolutive playlisting, advanced search features, multi-entry soundalikes, artists’ similarities/influences network as well as innovative discovery and navigation tools)
  • achieve psych-emotional profiling and matching (users, playlists, brands, etc.)


feature1