A Context-Aware View Over Social Media
The capabilities of mobile devices have increased with leaps and bounds. Nowadays, with a smartphone in hand, mobile users can run powerful mobile applications (e.g., office apps, games, route planners) anywhere and anytime. In addition to increased memory and processing power, mobile devices are also being outfitted with a range of detection technologies (e.g., NFC, GPS). Importantly, these technologies enable the collection of the mobile user's context. For example, at a short distance, technologies such as Bluetooth, RFID and NFC may discover surrounding objects, including people, places and things. In addition, GPS technology allows us to determine the location of user, and built-in accelerometers and digital compasses identify human motion and orientation. When equipped with these technologies, mobile devices are able to sketch an overall contextual picture about the mobile user.
At the same time, the explosive growth of social media allows people to connect each other in the online community. A wide variety of social media platforms exist to suit all sorts of social needs, such as communication in general (e.g., Facebook, Twitter), business and professional networking (LinkedIn), education (StudiVZ), music (Last.fm), and so on. In line with the current popularity of mobile devices, social media is being more and more deployed on mobile platforms. This includes mobile apps to supply instant and real-time access to social media (e.g., Twitter, Facebook), and directly share photos and videos online with family and friends (e.g., Instagram). Due to this mobile deployment, an excellent opportunity exists to automatically enhance social media interaction by leveraging captured mobile user context. Current examples include the geo-tagging of pictures shared on a social media platform (e.g., Facebook, Google Plus, etc.), and automatically discover services based on the user's location (e.g., Foursquare).
In particular, an important opportunity exists to realize a full, bi-directional integration of social media with the physical world. In one direction, context collected by mobile devices could be utilized to automatically enrich social media content. For example, information of a new place that users visited could be shared to the online community automatically. This automation reduces the burden of manual user's context annotation. In another direction, the relevant and useful information can be automatically retrieved from social media, by leveraging the mobile user's context. For example, when travelling to a new place, any information related to that place obtained from social media (e.g., the user's friends have been there, or information of some event is occurring at that place) could be interesting or helpful to the user.
The goal of this thesis is to investigate and develop a generic software framework, realizing a full bi-directional integration of social media and real world activities. Furthermore, it allows integrating any social network, and thus provides us an opportunity to exploit context-relevant information from different sources at once. In addition, integrating multiple social networks is necessary because people are typically subscribed to different social network platforms for different purposes. Finally, an android application is built upon this framework. This is a proof-of-concept application with a user-friendly mobile interface, grouped around important social media concepts.