Yet that’s exactly the direction taken by ViewChange.org , a project of Link Media’s LinkTV nationwide news, events and culture television channel and website. With funding from the Bill and Melinda Gates foundation, ViewChange.org uses video to tell stories about progress in global development, with an eye to influencing action on the part of everyone from average citizens to the media to policymakers to NGO staffers. Non-profit organizations, film distributors and individual filmmakers contribute to its own videos their documentaries, news reports, and other films to the site, and as users click through video links in featured topics, from water and sanitation to governance and transparency to environment, the site uses semantic web technologies to dynamically generate links from the video being viewed to related video media, news stories, topics and actions – links that change based on where you are in the video.
Not that you would know that semantics are behind the scene. And that’s a good thing (see The Semantic Web Blog’s recent stories that include reflections on the importance of designing from the user perspective, here and here). “We put a lot of effort into making the site approachable,” says Hannah Eaves, ViewChange project lead. “It would be easy to have this heavy global development content in a package that looks like every other web site out there.” The subject area is serious, but the inviting design and semantic web technologies serve bigger purposes: Making the heavy topics it engages with more accessible and making it easy for users to find what they’re looking for when they type it into the search bar – and then share those finds with their communities, especially with those members who may have a hand in determining where organizational or governmental global development funding dollars are spent.
And even though the semantic technologies powering this are invisible to the user, they’re important to furthering ViewChange.org’s mission. “The web technologies being used underneath in markup – you don’t see that as the user but we hope that gets picked up by search engines and other sites, so that the content on ViewChange is more easily discovered and accessible by other sites,” says Rob DiCiuccio, ViewChange Technical Lead at the site’s design and development services partner, Definition Studio. Brand experience agency Method created the front-end designs.
Semantic Tech Steps In
So, just what are some of those under-the-hood semantic technologies and models being leveraged by Viewchange.org? Having its heritage in the TV broadcast space, one bit of good news is that it’s almost par for the course for text to be created around videos, and about 80 percent of the external producers it works with have some sort of transcript to give ViewChange, for its semantic technologies to use for help in extracting entities, discovering related content and integrating with the world of Linked Data. All told some ten different APIs are mashed together in the project, including the Zemanta API that is used for entity extraction and analysis. “The reason we chose Zemanta was because of the amount of Linked Data identifiers it returns,” explains DiCiuccio. “It’s very good at identifying people, places, and topics, but the biggest reason was that it returned Freebase and DBpedia URLs with pretty much all of its entities, which we are using heavily.” Once ViewChange.org has the DBpedia and Freebase identifiers, with all their additional metadata on persons, places and things, it can go out at any point and pull additional data to further classify topics. “We can add layers of additional metadata because we have those links created,” he says.
ViewChange tested other solutions, such as OpenCalais, which it found to work very well with structured articles. But, Eaves says, Zemanta is better at dealing with more informal language, which may be why it performs well with spoken-word transcripts. ViewChange also is contributing back to the Linked Data community with its own entity entries for items that haven’t yet penetrated the space, whether it be one of the hundreds of languages in Africa that are little known to the world or the name of a minister in some small country without its own online records. OWL:sameAs as and Friend of a Friend (FOAF) are used for relating videos to topics.
Facebook’s Open Graph is represented, as well, in the service of the mission of sharing content and furthering its discovery. “We want people to watch a video and share it with their communities online and social networks, and in that sense Open Graph lets us tell Facebook that this is the title of the video, the description, the image for the post and that sort of thing,” DiCiuccio says. Eaves adds that there could be more ways to leverage the protocol in the future. “I love the idea of facilitating topic-level conversations across communities. So if a video is identified as being about a certain topic and others in your Facebook community have interest in that topic, maybe there’s some way in the future [to use Open Graph] to have conversations around a topic area,” she says. “It’s kind of an investment for maybe what’s coming.”
Some of the main technologies represented in ViewChange soon will be making their way into a prototype new site with international video at its core that’s got some support from the Knight Foundation and will be released as an open source project later this year on Git Hub. “LinkTV’s international news program will be the centerpiece and all sorts of news will be coming in from different providers,” says Eaves. Some video on ViewChange.org currently streams off of YouTube using ViewChange’s advanced embedded player for adding metadata, and that will be a feature of the new site too, opening the door to relating clips to a lot of different content libraries.