Context-aware Collaborative Creation of Semantic Points of Interest as Linked Data

The widespread adoption of powerful mobile devices with permanent internet connectivity has made the decade-old visions of ubiquitous and pervasive computing reality. This "shift of computing platforms" from desktop to mobile particularly favored the development of location-based services, which are aware of the user's geographic location and adapt their behavior accordingly.

 

In this kind of system, points of interest (POIs), which are the most basic form of location-based content, provide useful information about specific geographic places. Whether users are able to quickly find the points of interest they need in a given situation, strongly depends on the quality of the POIs' annotations. Unstructured text descriptions are harder to interpret for machines than folksonomic tags, which in turn are less machine-readable than formal ontologies with rich semantics. The consequence is a significant advantage of semantic points of interest over alternative forms of categorization.

 

The aforementioned shift toward mobility coincided with another socio-technological development, which is also strongly related to location-based services: the remarkable growth of user-generated content on the Web in the context of the Web 2.0. The traditionally well-defined boundaries between consumers and producers have blurred and loose collaboration of random volunteers has enabled the creation of content not only of acceptable quality, but especially for areas on the far end of the long tail that had not been covered before. User-generated content comprises a variety of different types of content, including the prominent example Wikipedia as well as geographic projects such as OpenStreetMap and other constituents of the emerging field of "neogeography".

 

This thesis proposes a system which enables its users to collaboratively create, share, and modify semantic points of interest as well as to adapt the underlying ontology of annotations. The users interact with the system through context-aware mobile clients, which communicate with a server over the Internet. The csxPOI system integrates both the developments of ubiquitous computing and extensive volunteer collaboration to create a geographic dataset with explicit, emergent, semantics. In the spirit of the Semantic Web, all permanently stored data is expressed in RDF and made publicly available according to Linked Data principles in order to advocate its system-external use.