Abstract
In a variety of applications relying on geospatial data, getting insights into heterogeneous geodata sources is crucial for decision making, but often challenging. The reason is that it typically requires combining information coming from different sources via data integration techniques, and then making sense out of the combined data via sophisticated analysis methods. To address this challenge we rely on two well-established research areas: data integration and geovisual analytics, and propose to adopt an ontology-based approach to decouple the challenges of data access and analytics. Our framework consists of two modules centered around an ontology: (1) an ontology-based data integration (OBDI) module, in which mappings specify the relationship between the underlying data and a domain ontology; (2) a geovisual analytics (GeoVA) module, designed for the exploration of the integrated data, by explicitly making use of standard ontologies. In this framework, ontologies play a central role by providing a coherent view over the heterogeneous data, and by acting as a mediator for visual analysis tasks. We test our framework in a scenario for the investigation of the spatiotemporal patterns of meteorological and traffic data from several open data sources. Initial studies show that our approach is feasible for the exploration and understanding of heterogeneous geospatial data.
Highlights
In a variety of applications relying on geospatial data, getting insights into heterogeneous geodata sources for decision making is crucial and has attracted a lot of attention
When the geologists analyze such data, they have to know that in order to get all the meaningful wellbores, they need to look into the table WELLBORE and, more importantly, the value of a specific column, namely REF_EXISTENCE_KIND, has to be ‘actual’, but not any other value [6]. Such knowledge is normally only known to the IT management team but not to the geologists. We argue that this challenge of the semantic gap is not merely an engineering problem, it has to do with how geo-concepts and raw data sets are related, and addressing it in an adequate and general way requires both new methodologies and techniques
We propose an ontology-based framework (called Ontology-based Geodata Integration for Geovisual Analytics (GOdIVA)) for data integration and analysis, which consists of two modules centered around an ontology: (1) an ontology-based data integration (OBDI) module, in which mappings specify the relationship between the underlying data and a domain ontology; (2) a geovisual analytics (GeoVA) module, designed for the exploration of the integrated data, by explicitly making use of standard ontologies that are defined by standard organizations, or are de-facto standards used in certain domains
Summary
In a variety of applications relying on geospatial data, getting insights into heterogeneous geodata sources for decision making is crucial and has attracted a lot of attention. While most existing visual analytical systems are only able to deal with geospatial big data of particular types, developing visual analytical approaches that can directly apply visualization methods to data of different formats, representing different kinds of phenomena, is still challenging This problem is acknowledged in the geovisual analytics community [3], where it is stressed that variety is one of the key issues in the research agenda related to visual analytics in the context of geospatial big data. In reality, geoscientists have to deal with data structured and organized in possibly many different ways, so that the correspondence between these data and the core concepts of interest to the scientists is often unclear Making this correspondence explicit requires a huge effort in finding the appropriate data, and in cleaning and preparing it for analysis.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
