What Spatial Metadata to Use & Why?
Work has begun to make a subset of Landmap spatial metadata more linkable and searchable. There have been discussions and an exchange of ideas within Mimas as well as external academic institutions to identify which datasets would be most useful as linked data. Landmap have decided to use:
- Thermal Imagery
- Building Class
- Building Heights
The above data are one of the smallest in number within the Landmap data archive but are of great interest for a vast variety of users. These data link very well with socio-economic data and other datasets from Edina.
The Pathway to Linked Data Conversion
These ISO 19115/19139 geospatial metadata records are represented in XML and to publish metadata in the Semantic Web requires mapping from the metadata schema to the RDF data model.
There has been exchange of ideas to create a Landmap data model diagram and carry out the crosswalks.
UML Diagram for ISO19115
This provides the high level UML representation of Landmap geospatial metadata currently in XML.
Landmap Data Model Diagram
Next steps are to do an XSL transformation to produce RDF files. We have plans to look into other options for producing the RDF files such as using the Java based framework called JENA.
Progress has been made on deciding the vocabularies using the standard OWL available for the ISO 19115 metadata standard. The three main vocabularies we shall be considering are Dublin Core Terms (dct/dcterms), Data Catalog (dcat) and Friend-Of-A-Friend (foaf).
Geographic metadata crosswalks have to be used to implement mapping from geographic metadata schemas to the RDF data model. The crosswalk log is added as metadata description and is transformed into the RDF data model where each metadata term is expressed as RDF. The below table shows a sub sample of the crosswalk:
Landmap has been participating in the discussions within Mimas to develop a triplestore database which would store linked data for all Mimas services. Once the RDF is stored in the database Landmap will be working towards publishing the linked data to the web.
Wider Work & Discussions
Landmap has been communicating with members of staff at Edina to facilitate the sharing of knowledge and experience regarding conversion of geospatial metadata to linked data. In addition there has been communication with academics outside of the UK e.g. University of Zaragoza.