Scottish Independence Referendum on Twitter

We present a new tool that monitors discussions on Twitter about the Scottish Independence Referendum.

Sensing User Generated Input for Improved Media Discovery and Experience

SocialSensor is developing a new framework to enable real-time multimedia indexing and search from the online social networks. The European Union funded FP7 project moves beyond conventional text-based indexing and retrieval models by mining and aggregating user inputs and content over multiple social media sites.

Social indexing will incorporate information about the structure and activity of the users' social network directly into the multimedia analysis and search process.

The SocialSensor News prototype is a tool designed for both news professionals, such as journalists and editors, and for casual news readers.

The system identifies key influencers in social networks who are active in sharing news or have demonstrated expertise in a particular field of interest. Information and data from these "news hounds" ensure that topics, trends and trust scores are useful for professional and general news users alike. Interfaces have been developed for web, mobile and tablet.

RGUs role

Our key contributions to this project include:

Topic detection - we developed novel methods to discover and track news stories as they break and spread across social networks. Our "BNgram" method finds words or phrases that show a sudden increase in popularity, which typically shows a response to a real-world event. Our software can then combine these messages to form coherent topics, complete with a human-readable label as a headline.

News hounds - we developed novel methods to automatically identify active and influential users of social networks with an interest in sharing the news. This includes journalists, bloggers, politicians and other opinion formers. Such users often forward up-to-date messages from eye-witnesses of major events, ensuring that a wide range of current information is available.

Context aware search - we have researched a number of methods to enhance search by taking account of the wider context that the user is in.

Evaluation - we have been active in the systematic evaluation of all aspects of the system, including user-centred `simulated task' studies and organizing an international evaluation competition in the area of news-focussed topic detection.

Current work

We have recently developed a system that monitors discussions on Twitter about the Scottish Independence referendum.


The RGU team:

  • Ayse Goker

    Ayse Goker

    Professor Ayse Goker is an expert in user-centred and context-aware information retrieval. She is theme leader in Digital Technologies at the IDEAS Research Institute and RGUs Principal Investigator for the SocialSensor project.  

  • Malcolm Clark

    Malcolm Clark Profile

    Malcolm Clark recently completed his PhD, studying structured text retrieval with an emphasis on genre. He also carried out user-centred evaluation for the SocialSensor project.  

  • Michael Heron

    Michael Heron

    Dr Heron is a lecturer in the School of Computing and Digital Media. He also developed an interface for the SocialSensor project's Scottish independence Twitter monitoring tool.  

  • Publications:

    1. Kompatsiaris, I.; Diplaris, S.; Corney, D.; Heise, N.; Klusch, M.; Jaho, E.; Geurts, J.; Liu, Y.; Petkos, G.; Papadopoulos, S.; Sarris, N.; Goker, A.; Spangenberg, J. (2014) "Real-Time Social Media Indexing and Search", Proc. International Broadcasting Convention (IBC), 12-16 September, Amsterdam, Netherlands.
    2. Corney, D., Martin, C., Göker, A. (2014) "Spot the ball: Detecting sports events on Twitter", ECIR2014 (Advances in Information Retrieval), Amsterdam, Holland, Apr. 2014, pp. 449-454. Pre-print
    3. Corney, D., Martin, C., Göker, A. (2014) "Two sides to every story: Subjective event summarization of sports events using Twitter," ICMR2014 1st Workshop on Social Multimedia and Storytelling, Glasgow, UK, Apr. 2014. Pre-print.
    4. Papadopoulos, S., Corney, D. and Aiello, L. (2014) "SNOW 2014 data challenge: Assessing the performance of news topic detection methods in social media" in Proceedings of the SNOW 2014 Data Challenge, CEUR Workshop Proceedings Vol-1150, pp 1-8.
    5. Martin, C. and Göker, A. (2014) "Real-time Topic Detection with Bursty N-grams", Proceedings of the SNOW 2014 Data Challenge, Seoul, Korea, April 8, 2014, pp 9-16.
    6. Schifferes, S., Newman, N., Thurman, N., Corney, D.P.A., Goker, A. and Martin C. (2014) "Identifying and verifying news through social media: Developing a user-centred tool for professional journalists," Digital Journalism.
    7. Martin, C., Corney, D., Goker, A., and MacFarlane, A. (2013) "Mining Newsworthy Topics from Social Media", BCS SGAI Workshop on Social Media Analysis, Cambridge, December 2013. Full proceedings.
    8. Corney, D., Martin, C., Göker, A., Spyromitros-Xioufis, E., Papadopoulos, S., Kompatsiaris, Y., Aiello, L., and Thomee, B., "SocialSensor: Finding diverse images at MediaEval 2013," in MediaEval Workshop, Barcelona, Spain, Oct. 2013.
    9. Martin, C., Corney, D.P.A., Göker, A. (2013) "Finding newsworthy topics on Twitter", IEEE Computer Society Special Technical Community on Social Networking E-Letter, vol. 1, no. 3, September 2013.Aiello, L.M., Petkos,G. 
    10. Aiello, L., Petkos, G., Martin, C., Corney, D., Papadopoulos, S., Skraba, R., Göker, A., Kompatsiaris, I., Jaimes, A. (2013). "Sensing trending topics in Twitter." IEEE Transactions on Multimedia, 15(6)

    Social Sensor Logo