Semantic Stream Processing of Environmental Data / von Peter Wetz
Weitere Titel
Semantic Stream Processing of Environmental Data
VerfasserWetz, Peter
Begutachter / BegutachterinTjoa, A Min
ErschienenWien 2016
Umfang177 Seiten
HochschulschriftTechnische Universität Wien, Univ., Dissertation, 2016
Arbeit an der Bibliothek noch nicht eingelangt - Daten nicht geprueft
Abweichender Titel nach Übersetzung der Verfasserin/des Verfassers
Schlagwörter (EN)stream processing / semantic web / linked data / RDF / OWL / mashups
URNurn:nbn:at:at-ubtuw:1-6319 Persistent Identifier (URN)
 Das Werk ist frei verfügbar
Semantic Stream Processing of Environmental Data [13.41 mb]
Zusammenfassung (Englisch)

Whether we cope successfully or fail to deal with the world's environmental challenges will be determined in cities where, since 2008, more than half of the global population resides. Recently, also the application of computer science methods to solve environmental issues is increasingly promising. In this thesis we present an approach to enable citizens to make well-informed real time decisions based on environmental data. To this end, we leverage semantic web technologies as a practical means to overcome the obstacles of (i) environmental data integration, (ii) identifying data stream management engines to process real time environmental data, and (iii) enabling ecient use of environmental data streams for city stakeholders. We develop an ontology-based approach to integrate highly heterogeneous and dynamic environmental data sources. We present a novel vocabulary that combines and extends two de-facto standard vocabularies, that is, the Semantic Sensor Network Ontology and the RDF Data Cube Vocabulary. Further, we create a framework to evaluate suitable RDF Stream Processing (RSP) engines based on the special requirements of the environmental data domain, such as processing of high-frequency data, providing correct results, and scalability. This framework called YABench facilitates the identication of appropriate RSP engines under varying circumstances for scenarios in the environmental domain. After we identify C-SPARQL as a suitable RSP engine, we propose Linked Streaming Widgets. Linked Streaming Widgets represent lightweight semantic modules based on stream data, which can be combined to web applications by end users. By doing so, users can author their own mashups integrating environmental stream data sources, ultimately supporting well-informed decision making. We implement this concept as an extension of a mashup platform. To demonstrate its feasibility, we present and discuss two use cases based on citybike and air quality data, respectively, and perform performance evaluations indicating the practicability of Linked Streaming Widgets.