I’ve been trying to capture the richness and variety of relationships within information about music. Noel at CarbonIQ started a brief discussion on music maps, which highlighted some really interesting ways of linking information around music, but didn’t develop that far beyond well-known, or slight, examples. Obviously Amazon and Allmusic have the basics down pat – simple catalogue-related facets of a release, which make it easy to find and sell music.
However, a cursory glance at these two bits of ephemera indicate there are other ways of thinking about music: Jeremy Deller’s nice maps of the connections between Acid House and Brass Bands (from his Acid Brass release); and sixties Zappa-protegé Wild Man Fischer’s own unique view of the musical universe.
So, how can we develop "Like this? Try this!" beyond simple catalogue relationships?
As we do more user research into how people think about and discover music, we’ll find lots of angles i.e. people getting into classical music via film soundtracks; discovering jazz whilst standing in Coffee Republic; reading about Vienna and developing a taste for both the Second Viennese school and ‘glitch’; music for adverts; etc.
Not one to miss a trick, Amazon extend this with their user-generated listmania, enabling people to group music together around their own relationships, like Essential Tortured Troubadours, or even music for drugs, love, suicide or enlightenment.
So, taking all this on, whilst engaged in a redesign of a large music site, I constructed a little diagram of basic facets around music, which I’m sharing with you in the spirit of open source concept development(!). I sketched this up to introduce some new angles into the architecture of the site; to begin to think about faceted classification(s) around music-related content, in terms of helping people discover new music:
I like to think of it as a 6-dimensional diagram, here presented in 2-D (use your imagination). The facets are:
Artists
The most basic, and one covered nicely by Amazon et al. Miles Davis is related to John Coltrane because they played together. I’m also using this as shorthand for basic catalogue-oriented relationships, such as song, composer, album etc. (though it’s worth noting there’s some interesting possible relationships here too, and some classic visualisation methods).
Genre
Another basic relationship: Miles Davis and John Coltrane both played jazz. Used by many music sites, though sparingly by Amazon who employ collaborative filtering most of the time (save Classical, which has more complex catalogue-based relationships of performers, conductors, composers etc.). I’ve previously discussed how record shops use genres, and how products like Echo can provide clustering beyond genres.
Time
Grouping music around time i.e. Miles Davis and John Coltrane are both most closely identified with the late-50s thru mid-60s. This goes beyond providing a simple timeline by providing a rich axis of organisation beyond genre and even artform, whereby we can talk about a richer set of music from, say, the late-50s, and link to/from other related cultural/historical content from that time. From there, it’s a short step to Music for Paintings, or Music of the Atomic Age.
Location
As with Time, location can provide fabulously rich cross-genre, multi-artform, historical relationships e.g. Miles Davis and John Coltrane are both closely associated with New York. In music, we could make connections with Talking Heads, Grandmaster Flash, DNA, Studio 54, Edgar Varése, Steve Reich, Randy Newman, Tin Pan Alley or even The bloody Strokes. There are some hugely interesting possibilities, not least in old chestnuts like Manchester, Vienna, Canterbury, Chicago, Liverpool, Paris, Detroit, or Seattle, but, zooming into the world map a little, Wellington NZ, Eel Pie Island, Athens Georgia, Laurel Canyon, Oslo, or your hometown. In previous jobs, I’ve built products along these lines.
Esoterica/theme
Made-up word alert, but this alludes to one of the more interesting angles, that of the arcane connections that crop up around music, often in pub conversations! e.g. Did you know that Miles Davis painted, as did Joni Mitchell, Schönberg and Captain Beefheart? This is a hugely rich area, as it veers towards the less tangible, more contentious, and more entertaining e.g. "Great ‘second-string’ guitarists in early-70s US rock bands" (Dickie Betts, Bob Weir, Frank Sampedro, Denny Dias …). Well, quite entertaining 😉 ‘Theme’ can include artists who had records banned by the Beeb, great composers with beards, bands with names lifted from books or films etc. I mean really etcetera!
Technique/instrument
A relationship which is more relevant in ‘educational’ contexts, perhaps e.g. "Like Hendrix? Want to play like Hendrix?" (as if). But important too – enabling a linking around turntables, or clarinets, say.
Plotting the facets for, say, a biography of Miles Davis gives this:
There are a basic set to begin with, and obviously, ‘theme’ has many relationships nested within it. But they enabled me to begin to introduce some more interesting navigational angles into the new BBC Music site I’ve been working on (will be launched end-May). There you’ll see a profile of an artist, say Miles Davis, with link relationships organised around time and location, as well as artist and genre. It’s a small step but one with potential, I think. It’s already enabling the producers I’m working with to get in many more potentially useful links around artists, providing fruitful jumping-off points for users, avoiding pigeonholing, as well as creating new ‘incoming trajectories’ on objects (could be important long-term, given all the current talk around developing 2-way+ relationships).
What I’d like to do next is develop these facets more coherently, and add in the missing links, and then develop ways of visualising these, which could be in the simple but powerful HTML link space, but perhaps more interestingly, new methods of data visualisation such as those employed by Martin Wattenburg for the History Wired or Idea Line. Or in Ben Schneiderman’s work. Something which approaches the layering and complexity of Textarc’s analysis, but for music?
Does anyone have other interesting facets for music-related information spaces to share/discuss, or thoughts about the visualisation of these relationships?
City of Sound 
Leave a comment