"Tags are pieces of information separate from, but related to, an object. In the practice of collaborative categorization using freely chosen keywords, tags are descriptors that individuals assign to objects."
Quoted from Wikipedia
Tag Patterns focuses on patterns that emerge when watching tags extracted from various popular social bookmarking websites. To study these "tag patterns" we sample tag data hourly. The collected tag data is based upon a set of static base tags:
When each base tag is queried the set of response tags are collected and then each response tag is queried in turn. By querying for tag data at this second level of depth we are able to effectively grow our database of inter-related tags.
This website provides visualizations, in both HTML and image form (coming soon), for the "tag pattern" data that we've collected. Future forward expect a public API which will expose portions of the data we've collected.
Update! As of May 17, 2008, 12:12 am Tag Patterns is tracking 167,807 total tags with 0 tag relationships.