Advanced

Transformation and linking of authoritative multi-scale geodata for the Semantic Web : a case study of Swedish national building data sets

Eiðsson, Eiður LU (2018) In Master Thesis in Geographical Information Science GISM01 20172
Dept of Physical Geography and Ecosystem Science
Abstract
Semantic Web technology has attracted much interest for its ability to integrate and deliver information from different sources, and increasingly, geodata are being published as triple data according to the graph structure laid out in the Resource Description Framework (RDF). Although the Semantic Web already contains a large volume of geotagged information, there is still a lack of suitable geometric representations for proper visualization thereof. Data held by national mapping agencies and other authoritative bodies would therefore be an invaluable addition, since they are typically of high quality and often produced at multiple scales. This thesis presents a case study where the feasibility of bringing authoritative multi-scale geodata... (More)
Semantic Web technology has attracted much interest for its ability to integrate and deliver information from different sources, and increasingly, geodata are being published as triple data according to the graph structure laid out in the Resource Description Framework (RDF). Although the Semantic Web already contains a large volume of geotagged information, there is still a lack of suitable geometric representations for proper visualization thereof. Data held by national mapping agencies and other authoritative bodies would therefore be an invaluable addition, since they are typically of high quality and often produced at multiple scales. This thesis presents a case study where the feasibility of bringing authoritative multi-scale geodata to the Semantic Web is investigated through a practical implementation involving nationally maintained Swedish building data sets. Three polygon-based products, each intended for display at a certain scale level, were transformed into RDF using a combination of GeoSPARQL and draft INSPIRE vocabularies. To test integration with other sources, the building data were linked with the community-maintained gazetteer GeoNames. The transformed and linked data were then uploaded to a triplestore, and from there, they were queried and visualized in a web map. The case study shows that the selected data sets can be brought to the Semantic Web for enhanced visualization of related information, but careful planning and harmonization of the original data should preferably precede the transformation process. Furthermore, they show that current geospatial vocabularies are largely suitable for the task, but will need some refinement and added capabilities before extensive data publication can commence. (Less)
Popular Abstract
The Semantic Web is seen as an extension of the Web, where information has well-defined meaning and can thus be more easily discovered and made sense of by both humans and machines. It is realized through a collection of technologies known as the Resource Description Framework (RDF), in which data are stored as subject-predicate-object triples in a graph structure. The triple data are written according to RDF vocabularies, which define concepts, relationships, and rules for specific domains. Among the main advantages of this approach is that it allows integration of data from different sources that would otherwise remain isolated. Its use for geodata has been the subject of several studies and projects in recent years, and shared (widely... (More)
The Semantic Web is seen as an extension of the Web, where information has well-defined meaning and can thus be more easily discovered and made sense of by both humans and machines. It is realized through a collection of technologies known as the Resource Description Framework (RDF), in which data are stored as subject-predicate-object triples in a graph structure. The triple data are written according to RDF vocabularies, which define concepts, relationships, and rules for specific domains. Among the main advantages of this approach is that it allows integration of data from different sources that would otherwise remain isolated. Its use for geodata has been the subject of several studies and projects in recent years, and shared (widely agreed upon) vocabularies are being developed for the geospatial domain. Since the way in which data on the Semantic Web are used cannot be fully predicted, geometric representations at multiple scales are required for proper visualization of linked information therein, something that can be provided by authoritative bodies such as national mapping agencies. The aim of this thesis is to assess the feasibility of bringing authoritative multi-scale geodata to the Semantic Web for enhanced visualization of related information. To this end, Swedish national building data sets at three scale levels were transformed into RDF triple data and linked with a community-maintained gazetteer. The transformation was done using GeoSPARQL, a fundamental geospatial vocabulary, and draft vocabularies based on the INSPIRE application schemas. The output was then stored in a triplestore (graph database) and tested through example queries within a purpose-built viewer. Although the multi-scale building data could be successfully transformed and linked, it is apparent that methods and vocabularies for this purpose need to be further developed and tested. (Less)
Please use this url to cite or link to this publication:
author
Eiðsson, Eiður LU
supervisor
organization
course
GISM01 20172
year
type
H2 - Master's Degree (Two Years)
subject
keywords
GIS, RDF, linked geodata, INSPIRE application schemas, data transformation, cartographic scale
publication/series
Master Thesis in Geographical Information Science
report number
90
language
English
id
8957696
date added to LUP
2018-09-03 16:11:34
date last changed
2018-09-03 16:11:34
@misc{8957696,
  abstract     = {Semantic Web technology has attracted much interest for its ability to integrate and deliver information from different sources, and increasingly, geodata are being published as triple data according to the graph structure laid out in the Resource Description Framework (RDF). Although the Semantic Web already contains a large volume of geotagged information, there is still a lack of suitable geometric representations for proper visualization thereof. Data held by national mapping agencies and other authoritative bodies would therefore be an invaluable addition, since they are typically of high quality and often produced at multiple scales. This thesis presents a case study where the feasibility of bringing authoritative multi-scale geodata to the Semantic Web is investigated through a practical implementation involving nationally maintained Swedish building data sets. Three polygon-based products, each intended for display at a certain scale level, were transformed into RDF using a combination of GeoSPARQL and draft INSPIRE vocabularies. To test integration with other sources, the building data were linked with the community-maintained gazetteer GeoNames. The transformed and linked data were then uploaded to a triplestore, and from there, they were queried and visualized in a web map. The case study shows that the selected data sets can be brought to the Semantic Web for enhanced visualization of related information, but careful planning and harmonization of the original data should preferably precede the transformation process. Furthermore, they show that current geospatial vocabularies are largely suitable for the task, but will need some refinement and added capabilities before extensive data publication can commence.},
  author       = {Eiðsson, Eiður},
  keyword      = {GIS,RDF,linked geodata,INSPIRE application schemas,data transformation,cartographic scale},
  language     = {eng},
  note         = {Student Paper},
  series       = {Master Thesis in Geographical Information Science},
  title        = {Transformation and linking of authoritative multi-scale geodata for the Semantic Web : a case study of Swedish national building data sets},
  year         = {2018},
}