p2pnet - rss feed: Project Unfound Artist: Merchanical Turk
p2pnet news | Freedom:- Bryan Hance runs RadioActivity.fm to help small-market stations stay SoundExchange-compliant.
That’s why Project Unfound Artist, launched by Fred Wilhelms and p2pnet, caught his eye.
The idea is to track down artists who haven’t been paid by ex-RIAA service SoundExchange because it can’t seem to find them.
“I’m interested in the whole process from a data-management perspective, partly because my service is helping feed this chunk of SoundExchange data (albeit in a very small way),” he says in an email, going on:
“I’m also a huge nerd and I wanted a reason to try out Amazon’s Merchanical Turk, and this seemed like a no-brainer.’
For anyone who’s never heard of the Amazon Merchanical Turk (and that includes me), it’s a, “one of the suite of Amazon Web Services, a crowdsourcing marketplace that enables computer programs to co-ordinate the use of human intelligence to perform tasks which computers are unable to do,” explains the Wikipedia.
“Requesters, the human beings that write these programs, are able to pose tasks known as HITs (Human Intelligence Tasks), such as choosing the best among several photographs of a storefront, writing product descriptions, or identifying performers on music CDs. Workers (called Providers in Mechanical Turk’s Terms of Service) can then browse among existing tasks and complete them for a monetary payment set by the Requester. To place HITs, the requesting programs use an open Application Programming Interface, or the somewhat limited Mturk Requester site.”
OK?
So, “I just ran a little experiment I want to share with you all,” says Bryan, continuing >>>
I just found 94 of these artists for $40. Process took about 4 hours, including this writeup.
The Setup:
——————————
I took the first 5 alphabetical listings (A-E) from the missing artists list, turned them into individual files, and posted this as a task on Amazon’s Mechanical Turk (mturk.com). Here’s a copy of the job I posted:
—————-snip————————-
Help find lost artists online
This is a sample list of many ‘missing’ recording artists. We are trying to ‘find’ them using the web.
http://fronteranetworks.com/se/missing_A.txt
http://fronteranetworks.com/se/missing_B.txt
http://fronteranetworks.com/se/missing_C.txt
http://fronteranetworks.com/se/missing_D.txt
http://fronteranetworks.com/se/missing_E.txt
All of these people are ‘lost’ - so I need you folks to ‘find’ them, using Google, MySpace, or whatever methods you want.
Just pick one of these lists at random and spend the specified time to “find” as many of these people as you can on that list.
“Finding” an artist means providing one (or more) of the following, in order of ‘best’ to ‘worst’:
- Finding a contact email
- Finding the artist’s or band’s website
- Finding URL for their Facebook or MySpace page
- Finding an email, website, or MySpace/Facebook for their **manager or recording label**. (You can often do this by finding a CD from the artist and looking it up on Amazon, CDDB, or MusicBrainz)
Compile the list of the ones you ‘find’ and send it to me as a text file with the ‘found’ information underneath their names. If you spend more than 5 minutes on someone and can’t find them, just write “Could not find any info for this entry” instead and move on.
—————-snip————————-
Finalizing the task: I set this up on Mechanical Turk for a maximum of 5 participants at $8 each, and fired this off at 11pm or so. I then went to sleep. I picked up the completed tasks in the morning, but they had all been completed in about 4 hours after I sent them off. So the turnaround’s pretty good.
The Results:
————————————————–
5 people completed these tasks:
Person 1: Processed 27 artists, found 27, missed 0.
Person 1 results: http://fronteranetworks.com/se/HIT1.txt
Person 2: Processed 24 artists, found 13, missed 11.
Person 2 results: http://fronteranetworks.com/se/HIT2.txt
Person 3: Processed 24 artists, found 20, missed 4.
Person 3 results: http://fronteranetworks.com/se/HIT3.txt
Person 4: Processed 24 artists, found 24, missed 0.
Person 4 results: http://fronteranetworks.com/se/HIT4.txt
Person 5: Processed 10 artists, found 10, missed 0.
Person 5 results: http://fronteranetworks.com/se/HIT5.txt
Total artists processed: 109
Total artists found: 94
Total artists missed: 15
Conclusions:
It’s easy to find these artists.
It’s fairly cheap to find these artists - I found them at ~42 cents per artist.
At this rate we could eat that whole 7,874 missing-artist list for about ~$3,300. (Anyone want to write me a grant?)
Also, as you can see from the results, some of these Turkers are fantastic - they provided contact emails, addresses, and multiple URL’s for some of the artists they found.
I haven’t contacted any of the artists on this list, because I’m not on Facebook or MySpace. So I’ll leave that to the readers here. If you do zip them a message, be sure to point them to this posting, and:
http://www.p2pnet.net/story/15761
http://www.p2pnet.net/story/15755
http://www.p2pnet.net/story/15752
So was Bryan happy with the results?
He was.
Excellent, Bryan. Thanks a lot 
Definitely stay tuned, and don’t forget to check out Project Unfound Artist.
Cheers!
Jon Newton - p2pnet
.
.Stumble It!
Use free p2pnet newsfeeds for your site. It’s really easy!
Subscribe
to p2pnet.net | | rss feed: http://p2pnet.net/p2p.rss | | Mobile - http://p2pnet.net/index-wml.phpNet access blocked by government restrictions? Use Psiphon from the Citizen Lab at the University of Toronto. Go here for details. Download here.
