Linked Science Enablement via Semantic Interoperability and Spatial Data Mining

Abstract

We are now witnessing a large-scale need for the use of spatial information. Examples range from monitoring of deforestation in the Amazon to everyday applications for navigation and map-based visualizations. However, the central theories for Geographic Information Science (GIScience) need to be developed further in order to support the range of useful applications of geographic information in the society. For this there is a need to understand whether the study of scientific assets and their spatial, temporal and thematic could help to reveal useful new theories. The task is to all of these assets like publications, scientific data, methods, tools or tutorials – and represent their links to each other and to space, time and themes. The core question thus is: can we interconnect all scientific assets? This calls for efficient methods to answer questions of where, when, what, who (and even why) about each asset. Linked Data provides means for both the representation and accessing of data about the scientific assets on the web. This way it becomes possible – likely for the first time – to study on a large scale what kind of stories the data about scientific assets has to tell. Spatial data mining together with ontological reasoning can help us make aggregations, visualizations, abstractions, and thus allow for exploration of massive collections of scientific data and related assets. If we achieve in interconnecting different assets then we can achieve Linked Science where not only different assets are connected but also different disciplines. In this paper we discuss the role spatial data mining, semantic interoperability, vocabularies and visualization to support enabling of Linked Science. We also provide examples from our different Linked Science projects to illustrate the ideas.