The widespread use of mobile devices for capturing people's daily activities and surroundings and the massive adoption of social networking and media sharing sites have turned the Web into a reflection of real-world events and activities. Therefore, mining online user-generated media content and data could be extremely valuable for better understanding and documenting the events and stories taking place around us. However, despite the great advances in data mining, machine learning, search and retrieval technologies, there is currently still a gap in enabling technologies for mining real-time streams of data and media content from social networks and for supporting efficient search and access to such content and to the derived insights. To this end, SocialSensor started in October 2011, and, being a three-year FP7 Integrated Project, aims at contributing to the advancement of such technologies. The core of this E-Letter aims at providing a succinct yet informative overview of the Project concepts, its main technologies and achievements.
The E-Letter starts with an overview paper on the Project by Diplaris et al., summarizing the project scope, its two use cases and the main challenges involved in pursuing its objectives. The next paper by Martin et al. discusses one of the most important problems dealt with by the Project: topic detection in Twitter. In particular, it describes a novel topic detection algorithm, developed by the project, to identify newsworthy topics in streams of tweets. Another interesting concept proposed by the project is the extraction of pleasant paths in a city based on user-generated content (Flickr photos). The paper Shortest Path to Happiness by Quercia et al. motivates and validates the proposed concept and briefly describes the developed approach. The next paper by Papadopoulos et al. focuses on the important task of social multimedia crawling and search, presenting a case study on social media content around the #OccupyGezi events. The last E-Letter paper by Cao and Klusch targets mobile settings, in which it is important to efficiently search for media content and information in a peer-to-peer setting, for instance in cases where large event attendants are interested in sharing media content with their peers without the mediation of a central mediation service.
Apart from these papers, there is a short overview of practical project outcomes, such as datasets and software that could be of value to the research community for building upon the project results. Furthermore, you will find the Chair's column, a featured article on sensing trending topics on Twitter, some featured blog articles as well as a list of CfPs from relevant events.
As this E-Letter is establishing itself step by step, we are looking for contributions from the research and industry fields. Don't hesitate to contact us if you are able to submit a single paper or even multiple on a certain topic or project for a coherent special feature. Look out for a Call for Papers being released soon!