Visualizing Usenet: a factor-analytic approach

Usenet is a widely popular system for organizing and distributing asynchronous discussions, but the newsgroup hierarchies and messages "threads" of Usenet do little to assist users in navigating through the hundreds of thousands of messages that appear on a single day. Messages may be cross-posted to newsgroups in different, even unrelated hierarchies, and threads wander off-topic as people follow up to prior messages. The social practices afforded by Usenet are not readily apparent in the interface of the typical newsreader. In this paper, I utilize corpus linguistic methods to suggest strategies for visualizing the underlying social practices of Usenet discussions. Using common factor analysis of word frequency lists extracted from a corpus of soc.culture Usenet messages. I present displays that are organized by language, topic of discussion, and frequency of interaction. Thus, by using information that is already present in Usenet messages themselves, new, potentially more meaningful displays of topic and social interaction can be obtained.