25 Feb Metadata Crowdmining
We need to tag 64,000 photographs with expert data.
Thousands of people are going to come and help us, some for hours at a time, and we will attract them through the pleasure of play. How can tagging an archive of old photographs ever be an enjoyable experience?
Contemporary designers are more and more mixing play with work, and at Tiltfactor, we’re interested in how this happens, and how one can foster expertise in the process. We’re also interested in how values manifest in these data-driven systems. In making games for the social good, and games that result in real-world contributions, the designer must rely on ideas beyond innovative game mechanics and good old-fashioned playtesting. The data entry system for repositories such as databases need to be so enjoyable, well-organized, and instantly rewarding that people approach the tasks with motivation playfully like… a game.
Crowdsourcing is already being done successfully in a few other games. Particulary inspiring for us is the work of Luis von Ahn and his “ESP Game” that is reportedly played by some users more than 40 hours a week! But there are obvious issues with crowdsourcing that such work suggests need fixing. Simple match mechanics inspire players to provide only the most common, obvious tags regarding a given image. The key problem is that archives are specific. They require specialized, precise information in order for their contents to remain relevant or useful to the people who wish to use them.
Professor Mary Flanagan has said that “Crowdsourcing is great for lowest common denominator data. But we also need to go beyond crowdsourcing and tap groups from each group’s base of expertise.” Our design team is thus challenged to encourage students, researchers and common users to be able to search for ridiculously specific items inside a given archive by traveling paths laid out by our players.
“Computer, show me happy pictures containing any species of duck in Florida before the year 1954.”
Our designs for our metadata research are presented within a small suite of enjoyable games that encourages players to think harder and to dig deeper into the depths of human knowledge. More on these designs to come.
~Tiltfactor Team Member Brendan S, who is using a mental-mapping program called x-mind that is insanely useful (and free!). Check it out at www.xmind.net.