Efficient Management for Geospatial and Temporal Data using Ontology-based Data Access Techniques

Μπερέτα Κωνσταντίνα

Unit:

Department of Informatics and Telecommunications
Πληροφορική

Deposit date:

2019-11-11

Year:

2019

Author:

Bereta Konstantina

Dissertation committee:

Εμμανουήλ Κουμπαράκης, Καθηγητής, Τμήμα Πληροφορικής και Τηλεπικοινωνιών, ΕΚΠΑ
Γιάννης Ιωαννίδης, Καθηγητής, Τμήμα Πληροφορικής και Τηλεπικοινωνιών, ΕΚΠΑ
Δημήτριος Γουνόπουλος, Καθηγητής,Τμήμα Πληροφορικής και Τηλεπικοινωνιών, ΕΚΠΑ
Νικόλαος Μαμουλής, Καθηγητής, Τμήμα Μηχανικών Η/Υ και Πληροφορικής, Πανεπιστήμιο Ιωαννίνων
Γιάννης Θεοδωρίδης, Καθηγητής, Τμήμα Πληροφορικής, Πανεπιστήμιο Πειραιά
Δημήτριος Πλεξουσάκης, Καθηγητής, Τμήμα Επιστήμης Υπολογιστών, Πανεπιστήμιο Κρήτης
Diego Calvanese, Καθηγητής, Τμήμα Επιστήμης Υπολογιστών, Πανεπιστήμιο Bolzano

Original Title:

Efficient Management for Geospatial and Temporal Data using Ontology-based Data Access Techniques

Languages:

English

Translated title:

Efficient Management for Geospatial and Temporal Data using Ontology-based Data Access Techniques

Summary:

The data model RDF and query language SPARQL have been widely used for the integration of data coming from different souces. Due to the increasing number of geospatial datasets that are being available as linked open data, a lot of effort focuses in the development of geospatial (and temporal, accordingly) extensions of the framework of RDF and SPARQL. Two highlights of these efforts are the query language GeoSPARQL, that is an OGC standard, and the framework of stRDF and stSPARQL. Both frameworks can be used for the representation and querying of linked geospatial data, and
stSPARQL also includes a temporal dimension.

Although a lot of geospatial (and some temporal) RDF stores started to emerge, converting geospatial data into RDF and then storing it into an RDF stores is not always best practice, especially when the data exists in a relational database that is fairly large and/or it gets updated frequently.

In this thesis, we propose an Ontology-based Data Access (OBDA) approach for accessing geospatial data stored in geospatial relational databases, using the OGC standard GeoSPARQL and R2RML or OBDA mappings. We introduce extensions to an existing SPARQL-to-SQL translation method to support GeoSPARQL features. We describe the implementation of our approach in the system Ontop-spatial, an extension of the OBDA system Ontop for creating virtual geospatial RDF graphs on top of geospatial relational databases. Ontop-spatial is the first geospatial OBDA system and outperforms
state-of-the-art geospatial RDF stores. We also show how to answer queries with temproal operators in the OBDA framework, by utilizing the framework stRDF and the query language stSPARQL which we extend with some new features. Next, we extend the data sources supported by Ontop-spatial going beyond relational database management systems, and we present our OBDA solutions for creating virtual RDF graphs on top of various web data sources (e.g., HTML tables, Web APIs) using ontologies and mappings.

We compared the performance of our approach with a related implementation and evaluation results showed that not only does Ontop-spatial support more functionalities (e.g., more data sources, more simple workflow), but it also achieves better performance. Last, we describe how the work described in this thesis is applied in real-world application scenarios.

Main subject category:

Technology - Computer science

Keywords:

Linked spatiotemporal data, Spatiotemporal databases, Semantic Web

Index:

Yes

Number of index pages:

Contains images:

Yes

Number of references:

Number of pages:

133