Wikipedia Scanner

I learned tonight (reading on Wired) about Wikipedia scanner which lets you look for Wikipedia posts by organization name. I understand Virgil Griffith (a Caltech grad student) hosts a database of Wikipedia edits by IP, date, topic, comments, etc., including the reverse DNS lookup info for all the IPs from which the edits originated. The site is set up so you can enter an organization name (e.g. Microsoft or Department of Energy) and see what topics someone from those organizations contributed to or commented on, and it then provides specific edits in Wikipedia "diff" form. It can be politically touchy- I guess that's in part why it was set up- to see what organizations are putting shameless spin on what topics. Wired hosts a crowd-sourced "salacious edits" collection that is fun to peruse or contribute to.

This is a great example of combining two distinct sources of public information to make the information much more useful and interesting. Equally interesting would be the Wikipedia scanner's own collection of what IPs and organizations are reviewing what Wikipedia posts...

Visualizing Wikipedia's change history

Yuriy's presentation on wikis reminded me of a visualization from IBM's Collaborative User Experience Research group called History Flow. This visualization shows how the many different authors of a wiki page collaborate. An example from Discover Magazine demonstrates how the visualization can be used to study the history of an article. I think this is a really interesting way to look at collaboration.

More examples can be found here. The images are most dramatic for articles that are somewhat controversial or prone to vandalism because these articles have a high number of edits and lots of deletions and reversions.

Semantic Web Enablers

While doing some research I learned about the semantic wiki extension to MediaWiki called Semantic MediaWiki, as well as DBpedia. Both appear to be key approaches to enabling the semantic web. Semantic MediaWiki, described in this publication, describes an approach to eliciting Wikipedia-style knowledge from a broad base of contributors using semantic web W3C standards that permits the factual imormation to be machine processable. DBpedia is a community effort to transform Wikipedia information into RDF triples, rendering Wikipedia knowledge usable for the Semantic Web. These two tools illustrate the spectrum of approaches to populating Semantic Web knowledge stores, ranging from manual to automated methodologies.