Managing Norwegian Geospatial Data through Virtual Knowledge Graphs

Abstract

This thesis examines the use of Virtual Knowledge Graphs (VKG) to manage and query geospatial data, focusing on evaluating the performance and functionality of three different Knowledge Graph (KG)systems: Virtuoso, Ontop, and GraphDB. We aim to assess their compliance with the OGC GeoSPARQL standard and measure their query execution performance over data acquired from the endpoint of the Norwegian Mapping Authority. Geospatial data is increasingly critical for various applications, from urban planning to environmental monitoring. Traditional relational databases often struggle with the complex nature and high volume of geospatial data. To address this, we explore the VKG approach, which can integrate data from diverse sources into a unified, flexible graph structure, enabling more efficient data management and querying. We designed ten queries to evaluate the systems, divided into two groups: basic SPARQL queries to test fundamental capabilities and advanced GeoSPARQL queries to assess geospatial functionalities. The performance evaluation focused on execution times, providing insights into the ability of the system to handle geospatial data. The results reveal that both Ontop and GraphDB outperform Virtuoso in most scenarios, with Ontop demonstrating superior capabilities in handling complex GeoSPARQL queries. This suggests that VKGs, offer significant advantages for geospatial data management. Additionally, the study addresses data quality issues encountered during the exploration of the data, such as inconsistencies and redundancy, and proposes solutions to improve data quality. By providing an evaluation of these systems and addressing data quality challenges, this thesis offers insights and practical recommendations for researchers and practitioners in the field of geospatial data management.