Harnessing the power of storytelling to change the world™

On Linked Data and

Posted on November 17th, 2010 by Rob DiCiuccio

One of the primary goals of has been to bring the concepts of the Semantic Web to video. When we began conceptualizing the ViewChange platform, the use of Linked Data and participation in the Linking Open Data (LOD) cloud were core requirements, as we wanted to provide both semantic meaning and the potential for deep exploration. Being a video-centric site, this presented a unique set of challenges.

The Web of text

HTML, XML, RDF, JavaScript, CSS. The foundation of the Web is text-based. While online video is becoming increasingly popular, and now accounts for a large percentage of Internet traffic, video is still somewhat a second-class citizen on the Web. This is beginning to change, as new open video standards emerge, but as it stands, we have to transcribe the video in order to process its content. Once we have text transcriptions of the videos, we are able to operate on a level playing field. Video transcripts are analyzed using Entity Extraction tools, and topics (disambiguated Named Entities) extracted from the raw text. The resulting topics are linked to online databases such as Freebase and DBpedia, which provide the connection to the LOD cloud.

Metadata & accessibility exposes semantic metadata in a number of ways. RDFa markup is used throughout the site to identify resources and the relationships between them. Popular ontologies, such as Dublin Core, FOAF, and Facebook’s Open Graph protocol, are used to provide standard, interoperable metadata, and enhanced accessibility. Topic and video pages also provide RDF/XML representations, for efficient machine processing.

Linked Data concepts

ViewChange addresses the the four concepts of Linked Data, outlined by Tim Berners-Lee, as follows:

1. Use URIs as names for things

All topic and video resources on have unique permalink URIs, which serve as strong identifiers.

2. Use HTTP URIs so that people can look up those names.

Topic and video pages are accessible via HTTP, naturally.

3. When someone looks up a URI, provide useful information, using the standards (RDF*, SPARQL)

When a video or topic URI is dereferenced, structured data is returned using either XHTML+RDFa or RDF/XML.


4. Include links to other URIs. so that they can discover more things. topics are semantically linked to entities in databases such as Freebase and DBpedia (using owl:sameAs), while videos expose links to a wide range of related content.

Feedback welcome

Much like the Semantic Web itself, the Linked Data implementation in is a work in progress. We are very interested in feedback from the community, so please let us know if there’s something we could be doing better.

Tags: , , ,

One Response to “On Linked Data and”

Hi there,

I like your approach to semantically annotate video content using RDF. Given that most (all?) of your videos have transcripts, I was wondering if you were aware of SemWebVid, an approach to automatically annotate videos which I have presented at this year’s ISWC2010 in Shanghai: There’s also an interactive demo of the application available at I believe the application could be easily relatively easily be adapted to your needs. The source code is licensed under a Creative Commons license: Let me know if you need more details.