Bandfi: Unofficial Myspace Music Explorer

Bandfi lets you listen to any band on Myspace, while browsing a ‘quilt’ of similar, recommended and random artists. As you listen, the quilt updates to include friends of the currently playing band, allowing you to explore other bands in their network.

Features

• Works for ANY band with a Myspace account
• Queues a band’s friends to listen to next
• Filters out non-music profiles
• No interruptions from advertising
• Quilt provides a random selection of bands to explore
• Random selection is tweaked so you get a mix of popular and obscure bands

How It Works

On the surface, Bandfi is a simple way to listen to a band’s latest tracks on Myspace. Simply type in their URL and you’re away. But behind the scenes, there’s a lot of magic happening. Every time you listen to a band, it sends out a ‘spider’ to their page to fetch their player, thumbnail and name. It also visits the pages of their friends, checks if they’re a band and if so, queues them up, ready for you to listen to next.

All this information is stored in a database, which grows organically as the site is used. Listening to one band will add that band AND all their friends. It began with just one band and has since grown to around 15,000. This is just from my own personal use! Having bands in a database not only allows things like faster loading and searching, it also allows you to explore the connections between bands. 15,000 is only a small fraction of the millions of bands on Myspace, but you still get some interesting insights into which bands are popular amongst their peers. Here’s some examples:

1st – Sonic Youth (146 appearances)
2nd – Deerhoof (106)
3rd – Diplo (104)
4th – Grizzly Bear (95)
5th - Wilco (86)
6th – Neil Young (82)
7th – MIA (77)
8th – Ariel Pink’s Haunted Graffiti (77)
9th – Justice (75)
10th – Flying Lotus (74)
300th – Mogwai (30)
1,000th – Bruce Springsteen (19)
10,000th – Why are we building such a big ship? (3)

As you can see, the top 10 differs slightly from what Billboard has to offer. And the numbers drop off pretty quickly. The influence of Sonic Youth is quite astounding. DOUBLE the 10th place band. Then again, it is Sonic Youth. These statistics are used when generating the ‘random’ quilt when the page first loads. There’s a lot of terrible bands on Myspace so you need a way to filter them out. Being in another band’s top friends is usually a good indicator. So the more times a band appear in the top friends, the more likely it is at being selected. It works surprisingly well at delivering a mix of popular and obscure bands.

Why?

Even though Myspace’s dominance is being challenged, it’s coverage is still unmatched. Irregardless of a band’s size, you can usually count on them having a few songs on Myspace. And these songs will usually be the bands most recent or popular songs. How many other services can boast that? Where else can you go to hear the latest demos of that awesome band you just saw at the pub?

Myspace is such an incredible resource and yet it’s going to waste. They are only just starting to aggregate music from across the site into some sort of service, but to be honest, it’s not very inspiring. They have an enormous amount of data about the social connections between bands, which could be used to create an amazing recommendation engine, yet they are doing nothing of the sort. Bandfi is an attempt to shake things up a little and create a compelling experience out Myspace Music.

Future Improvements

Like all my side-projects, Bandfi competes for my precious little spare time. Who knows if I’ll ever return to it, but if I do, here’s what I have in mind:

1. Scenes

Using the friendship data between bands, it’s possible to identify ‘scenes’ within the music community. That is, bands with a similar style and close proximity. You begin with one band and gradually find more by looking at their friends and then the friends of friends etc. A good feature would be allowing you to pick one band and have Bandfi automatically generate a quilt composed of bands from the same scene.

2. Playlists

At the moment, Bandfi plays each band’s playlist of songs, but this is fairly limited. There’s no way to queue one band after another or make a mix. However, Myspace has built-in playlist functionality and I’ve figured out a way to create playlists programatically. This means it would (theoretically) be possible to build automatic playlists. Add in some recommendation data and you have an interesting prospect. I say ‘theoretically’ because I haven’t explored this very much, but it’s something I plan to do when I find time.

If you have any questions, requests or thoughts to add, leave a comment!