Previous articles on the website describe how SAM creates the whole new experience for Smart TV users, by employing inter-connected devices, in order to turn media interaction from passive and one-way to proactive and interactive. People may now consume and comment on TV programs as well as rate related information in what we defined as “2nd Screen” devices. But how is users’ context information managed and analysed, in order to personalize and enrich their experience [1] in such an environment?

Context Awareness

Context-awareness handles the problem of the overabundance of content by creating a profile for every user according to their preferences and social behaviour, which in turn serves as a guide for the selection of the appropriate and relevant material [2]. The main objective of the Context Analysis in SAM is to build social communities based on user-consumed media (videos and interactive 2nd screen windows), behaviour and interests expressed while interacting with the system and other users. The core software of this work, Context Manager, is collecting users’ contextual information from Social Media (Facebook, Twitter) and SAM dynamic communities, as well as their interactions with 1st and 2nd screen.

For example, when a user logs in to SAM platform, background listeners record which videos she watches in full screen or which extra information she “likes” at 2nd screen, as well as all her comments posted over various social communities. This interaction history is useful to form sophisticated user profiles. In order to save all those information efficiently [3], a graph database (Neo4j) is being used [4]. Thus, all SAM users and media items (named “assets”) are saved in the form of graph nodes (dots), while the interactions between them are shown as edges (lines) connecting those nodes.

Context Management graph example

Context Management graph example

Having this efficient database model, graph analysis techniques and user behavioural models are applied over the aforementioned Context Management graph to properly assess the relevance of each media asset to every SAM user.

Personalization of Social TV experience

Assessing the relevance of assets for every user enables personalizing the Social TV experience for them. This is realized by TV program recommendations or by prioritizing assets in 1st and 2nd screen for a user whenever she logs in.

Smart TV program recommendations, image by

Smart TV program recommendations, image by

Therefore, the Context Analysis processes result into the following recommendation/ prioritization functionalities [5]:

  • Personalized recommendations of 1st screen assets (videos) to users, based on their relevance with all available videos
  • Personalized recommendations of 2nd screen windows hosting sources of information such as Wikipedia articles to users, based on their relevance calculated values, while watching a TV program on 1st screen.

The provided functionalities enrich the Social TV experience for every end-user, highlighting most relevant media and adapting the Smart TV environment according to her profile.

Interested in SAM Context Analysis?

SAM wiki page contains plenty of information related to the research conducted within SAM project and software components created for this purpose. Other than that, you can read SAM research publications referenced below or directly contact the NTUA/ICCS Distributed Knowledge and Media Systems Group to learn more about our Context Management and Personalization algorithms and research.




[1]         Andreas Menychtas, David Tomás, Marco Tiemann, Christina Santzaridou, Alexandros Psychas, Dimosthenis Kyriazis, Juan Vicente Vidagany Espert, Stuart Campbell: Dynamic Social and Media Content Syndication for Second Screen. IJVCSN 7(2): 50-69 (2015)

[2]         C. Santzaridou, A. Menychtas, A. Psychas and T. Varvarigou, “Context management and analysis for social TV platforms,” eChallenges e-2015 Conference, Vilnius, 2015, pp. 1-10

[3]         Fotis Aisopos, Angelos Valsamis, Alexandros Psychas, Andreas Menychtas and Theodora Varvarigou, “Efficient Context Management and Personalized User Recommendations in a Smart Social TV environment”, GECON2016 Conference, Athens, 2016

[4]         Neo4j: The World’s Leading Graph Database,

[5]         Angelos Valsamis, Alexandros Psychas, Fotis Aisopos, Andreas Menychtas and Theodora Varvarigou, “Second Screen User Profiling and Multi-level Smart Recommendations in the Social TV of SAM”, 9th International Workshop on Social and Personal Computing for Web-Supported Learning Communities, SPeL 2016, Rome, 2016

Also read about:

Evaluating User Engagement with the SAM Platform

Spoken Dialogue Capabilities in SAM

Socialising Around Media with Wi-Fi Aware™

How SAM provides an endless stream of relevant & rich content

Surfing an Ocean of Extended Content

Adaptation in Spoken Dialogue for Second Screen Interaction

and: SAM and Marvelous Metadata

For additional information about the technologies involved in the SAM project take a look at our public wiki page and/or  follow us via our Twitter channel.