Linked data

Linked Data is rather like putting multiple databases on the web and giving each item of interest a globally unique key that allows you to make links between different databases. Or a way of giving Web pages some of the characteristics we know from databases, such as merging, linking and querying information.

In the Linked Data world, these identifying keys are also dereferencable URLs, which can give more machine-readable information when fetched, such as what they are connected to, what sort of a thing they describe and what properties or attributes they have. The items of interest can be anything – people, documents, places, pictures, videos, anything that can be identified.

Why was this of interest to NoTube?

Linked Data was of interest to the project in several ways:

1) Helping people decide what to watch

Enriching TV content metadata with Linked Data enables interesting links to be made between programmes, and selecting the types of linked data connections that are made (for example, connections based on place and time period, such as ‘London in the 1970s’) allows useful content filters to be created for recommendations.

2) Helping people to find out more about a programme

Further, by using URLs to identify TV programmes, NoTube can unambiguously link programmes to related information from different sources around the Web using the Linked Data Cloud.

3) Helping people to have smarter conversations about TV programmes

One of the project’s Use Cases, “TV and the Social Web”, was concerned with future Social TV scenarios involving people interacting with each other online while watching TV. However, Web-based conversations, about TV or anything else, cannot be held without URLs that allow you to refer to the thing you’re talking about.

Global channel and programme identifiers are therefore crucial in a distributed environment. For example, users may share information about programmes, recommend programmes to each other or bookmark a broadcast from work to view at home in the evening. In all cases users must be able to refer to a programme or broadcast with a location-independent identifier.

What NoTube has done in this area

Much work has been done on semantically enriching content and user data with concepts from Open Linked Data repositories.

Based on the observation that people would often like to know more about a programme they are watching, in one early NoTube demo we used Linked Data to show related information about a programme being broadcast. The additional information, in this case from Wikipedia, was presented to the user on a second screen.

Early NoTube iPhone app

Early NoTube demo showing use of a smartphone as a companion device

Much of the technical work was concerned with creating the the infrastructure for supporting Social TV APIs, in particular by disambiguating specific programme episodes resolvable via unique IDs. Having this infrastructure in place allows for the development of demos that illustrate the use of APIs to TV.

Find out more: See the Things to read section of this site, and read a blog post about experimenting with Linked Data to improve the TV experience.

Who was involved? Vrije Universiteit in Amsterdam, BBC Research & DevelopmentPro-netics, and Semantic Technology Institute International, in collaboration with other project partners.