NoTube has produced a number of Web based demonstrators of project results.
N-Screen is a demonstration of the social TV and second screen work of NoTube.
N-Screen is a web application to help small groups of people decide what to watch. Users can can share programmes with one another in real time, and change the TV channel using drag and drop (try it [TV]). If they also connect to their Beancounter account (below) they get personalised recommendations of available TV content.
N-Screen is a flexible web-based demonstrator using HTML5 technologies to show a possible future of simple rather than a ‘smart’ TV where the smarts stay in the tablet or other hand-held device and the TV acts as a player. It has undergone extensive user testing (see blogposts here and here) and a few design iterations based on those: there are therefore various versions of it, including one using programmes from BBC and TED (player).
More detailed technical discussion can be found in this blogpost and there is a short summary of the work on the BBC R&D blog. The code is available from github.
Beancounter is a user profile manager, where you control all of your data.
You can attach your social web accounts - Twitter, Facebook and LastFM, and it generates a profile for you that can be visualised or used for recommendations. You can use your Beancounter profile with TV applications like N-Screen (above) to get personalised TV recommendations. Try it!
You can read about Beancounter focus group testing and user reactions to the Beancounter interface, and the code is available in github.
NoTube Portal
NoTube provides and makes use of a number of Web-based services to support its work on enabling the convergence of TV and the Web. These services can be more easily accessed and used in development through the NoTube portal.
NoTube services and APIs
Recommendation
- the Linked Data Pattern-based Recommender: uses Linked Data as background knowledge to recommend programs. The Beancounter provides the user profiles (derived from the Social Web accounts of a user). Background knowledge in DBpedia, Freebaseand LinkedMDB are used to find potentially interesting commonalities between items and user interests. Two example of Pattern-based recommendations (generated real-time) are available: one is for the documentary ‘Paul Merton looks at Alfred Hitchcock‘ and the other is for the cooking program ‘Big Cook Little Cook‘. The Pattern based Recommender can also be tested in a Web UI where autocomplete will help you find the program title.
Metadata Enrichment
- Lupedia is a Web UI and a REST service provided by Ontotext to enrich TV program titles and descriptions with named entities identified using concept URIs drawn from Linked Data sources.
Metadata Mapping
- IRT provides a number of Web services for mapping between TV metadata in the TV Anytime, BMF and PrestoSpace formats.
- BBC produced a CRID Resolver which maps between DVB CRIDs for programmes and BBC http URLs for programmes and vice-versa.
User profiling
- The Beancounter is available as open source (Apache 2 license).