Below is an embeddable widget which, like Twitter's search widget, provides a live view of a stream of tweets. However, this one allows you to create much "smarter" views by issuing SPARQL queries over an RDF data stream. Why is this cool? Well, the more data that can be gotten out of a tweet, the more sophisticated real-time filters you can create. Imagine a stream of tweets by people with low Erdös numbers, for example, or about upcoming albums by bands you like. Rather than developing an entirely new Twitter application, just write a new SPARQL query. The example below is connected to a SPARQL endpoint which contains the Semantic Web Conference Corpus, a DBpedia dump, special-purpose data for ISWC 2010, and a real time stream of tweets about the conference. Based on the authors and #hashtags of tweets, we can begin to construct some useful filters.

To use this demo:

• Select a filter from this list:
• Watch the tweets appear. Any subsequent tweets will appear as soon as they arrive.
• You can easily customize this widget for your own application, changing the query, the SPARQL endpoint, the style of the widget, etc. Have a look at the source code.

To see the SPARQL query for each filter, click on the "SPARQL powered" link in the lower right-hand corner. These queries are executed against pre-loaded static data and TwitLogic-generated streaming data in this AllegroGraph triple store.

Disclaimer: this widget is very new, so there may be bugs. I believe it works in Firefox, Chrome, Safari and Opera. It does not yet work in Internet Explorer. If you encounter any other issues, please let me know.