Flickr tags are by people categorizing their own photos (that is, publisher-generated categorization.) Not people categorizing things they've found or aren't theirs. This leads to radically different behavior. Given this is a librarian, the second case is much more relevant and interesting.
One analysis without understanding the space or how the subject differs from the desired outcome (that is, categorization) seems very embarassingly shallow.
(Yes, this annoyed me the first time I read it, a while ago.)
Les Orchard (who used to work for delicious.com) has a good response to this article. Part of his argument is that Flickr is a really bad place to start a tag analysis, since people generally only tag their own stuff (the big win for tags comes when lots of people provide their personal tags for a single entity, meaning you can look at the data in aggregate). http://decafbad.com/blog/2009/01/18/tags-do-work-for-me-at-l...
The findings basically show that people generally don’t use verbs when tagging and that the titles and descriptions they write are almost always a more useful definition of the thing being tagged (especially when searching across large datasets).
Flickr tags are by people categorizing their own photos (that is, publisher-generated categorization.) Not people categorizing things they've found or aren't theirs. This leads to radically different behavior. Given this is a librarian, the second case is much more relevant and interesting.
One analysis without understanding the space or how the subject differs from the desired outcome (that is, categorization) seems very embarassingly shallow.
(Yes, this annoyed me the first time I read it, a while ago.)