I recently had a guest appearance at the Semantic Web Gang podcast. Together with MIT’s David Karger and the regular “gang members”, we discussed interfaces to the Semantic Web. The discussion was really intense at points as the participants had very differentiated standpoints.
It was largely agreed that visualization is a key factor to show the benefits of the Semantic Web and to foster industry adoption. Some key discussions: Should we worry about applications first, or step back and consider visualization at this early stage? How many forms of visualizations are there for one ‘thing’? Who builds them? Where do you get data from? Should users worry about that? How are queries expressed?
The participants also agreed that there will be an industry of people that build visualization widgets for specific things, which I thought is a really interesting scenario.
I talked a bit about Marbles and DBpedia Mobile and pleaded for open environments where data published by anyone can be taken into account. (Podcast)
flickr™ wrappr
12Oct07DBpedia extracts structured information from Wikipedia and publishes it as RDF. This allows for incredible queries against Wikipedia data, such as Soccer player with tricot nr. 11, playing for a club having a stadium with >40.000 seats, born in a country with >10M inhabitants.
I recently added support for geo-coordinates to DBpedia and built the flickr™ wrappr, a proof-of-concept work that provides photos for a given Wikipedia article using geo-coordinates and multilingual labels. The benefit: It is a fairly efficient and accurate way for a machine to find a picture of something. It is now part of the W3C SWEO Linking Open Data community project.
I really like the pictures it finds for the Eiffel Tower:
Update (04/10/2008): ReadWriteWeb says "This is pure geek hotness", and Tim Berners-Lee says “This is a neat addition”!
Update: Mashup of the day at ProgrammableWeb (10/17/2007) and at Mashup Awards (01/10/2008)!
