more steve ... tagger prototype preliminary analysis

I've begun working with the data from the steve.museum prototype tagger, in preparation for a presentation at the ASIST SIG/CR workshop in Austin TX on November 4, 2006.The preliminary results, discussed in my paper [PDF preprint on-line] confirm the gap between terms provided by social tagging and those provided in professional museum documentation. We're seeing about 90% new terms in the top four, most-tagged works -- a number that really surprised me. (The blue terms in the charts are the ones that match museum documentation.)

It's not that one approach is right and the other wrong, just that there seem to be two entirely different discourses, two methods of looking at, and talking about art that seem to be almost orthogonal. There are very strong correlations in the terms supplied by different taggers, and these terms often aren't found in the museum's cataloguing.

It's got me thinking about the assumptions we come to museum documentation with, and the challenges in re-purposing information across platforms.

jt