Memeta: A Framework for Multi-Relational Analytics on the Blogosphere

The “memeta” project is developing a framework for studying the structure and content of the blogosphere. We are particularly interested in how metadata about blogs can be discovered, extracted and computed, and how this metadata can be modeled, represented and analyzed to provide new blog related services. Weblogs, or blogs, are web sites consisting of dated entries (posts) typically organized in reverse chronological order on a single web page. They have become an important new way to publish information, engage in discussions and form web communities. Due to their high influence and specialized publishing infrastructure the subset they constitute is popularly known as the blogosphere. The dialogue oriented nature of the blogosphere has manifested in the use of Comments and Trackbacks as mechanisms for managing conversations. Blogs are also highly dynamic and embedded in a community of both casual and regular readers, factors that have led to the adoption of new content representation models supported by metadata in several formats. Blogs are considered as “killer applications” for the RSS (RDF Site Summary) and Atom syndication formats which recommend an XML-based representation for machine consumption, representing information like title, date, summary, etc. for content (e.g., a blog post). This metadata is used by aggregators and search engines to provide new blog-related services. The popularity of blog aggregation has in turn encouraged the publishing of explicit author name and sometimes complete profile information using the Dublin Core and FriendOf-A-Friend metadata vocabularies. In addition, blogs truly exploit the use of tags and the folksonomies they create. Tags are explicit metadata labels associated with a post to provide a highly subjective topic description, and has been found useful for search, browsing and “buzz” analysis. Blogs have also popularized the concept of referrals. Bloggers often publish frequently visited and followed blogs (known as blog subscriptions or blogrolls) in well structured formats like OPML (Outline Processor Markup Language), creating a referral chain that enables interesting topical and community oriented web navigation capabilities.