DiscoBro is a simple Web application that makes it possible to discover data resources available while browsing the Web. It currently makes use of the Archive Hub and the MusicNet datasets, but can be extended with any linked dataset providing both a SPARQL endpoint and links to DBPedia.
As the short video below demonstrates, it is easy to use. It takes the form of a bookmarklet that can be dragged in the bookmark bar of the browser. While reading a page, the bookmarklet can be clicked to make appear on the side of it terms that correspond to resources mentioned in the text of the page. Each term links to the description of the corresponding resource in DBPedia. If the resource is described in one of the other datasets (i.e., MusicNet or ArchiveHub), links will appear in the list, to connect to the discovered resources.
In the video below, a simple example is shown that demonstrate how DiscoBro can be used to find and browse resources related to J.S. Bach, starting from a page containing his full biography (Note: I highly recommend to watch it in 720p HD and fullscreen, otherwise, it is a bit hard to see what’s going on). Musical genres as well as other artists are linked that also relate to resources in MusicNet, which can be further explored from within the browser. An interesting aspect of DiscoBro however is that it is not restricted to these two datasets (and so these two domains). As the amount of cultural resources being made available as linked data (behind SPARQL endpoints) increases, more and more sources can be integrated providing more links from pages regarding any domain, and relating to any source of data (as long as they can be linked to DBPedia). In other words, DiscoBro is more of a general framework than a specific application of these two datasets. It demonstrates how, based on a few simple requirements, data can be linked to the browsing experience of Web users, allowing for a serendipitous discovery of available resources.
DiscoBro is based on DBpedia Spotlight, which is a tool for Named Entity Recognition using DBpedia resources as output. (Warning: Sometimes DBpedia Spotlight goes down. If nothing shows up in the pannel on the right of the screen, it is probably what happened). DiscoBro also uses SPARQL through the sparql.js library in Javascript.
Links:
- DiscoBro page (where a URL can be entered and the bookmarklet obtained
- Source Code of DiscoBro, which will probably evolve a bit in the next few months (don’t hesitate to modify/adapt/contribute). Note however that you will need an API key from Kasabi to use the MusicNet SPARQL endpoint (to be inserted in the musicnet_endpoint.php file)
- Youtube video of an example of using DiscoBro