The Landmap Service provides spatial data to the UK HE and FE. Landmap spatial data is provided under several different licensing agreements which restrict access and usage to research and teaching only in the UK. However the spatial metadata can be made widely available without any restrictions as open source under Creative Commons licensing. Hence the reason why Landmap has decided to use a subset of spatial metadata (thermal and building class) for this project to be provided as Linked Data in the semantic geo-web world.
The value of producing spatial metadata as linked data for a user would be to enhance the “searchability” for information within the geospatial community and elsewhere.
Ideas from the Community
Prior to the start of this project the Landmap team attended various events to start understanding about linked data and the key concepts e.g.
1. One-day ISKO UK conference: LINKED DATA: 14 September 2010, VENUE: UCL
This event was attended by Bharti Gupta, the Applications Developer, where she developed her basic knowledge of the semantic web and linked data. Nigel Shadbolt who is working with the pioneer Sir Tim Berners-Lee on the initial Linked Data work for the government, talked about the potential of building a Web of Open Linked Data using Semantic Web Standards.
2. Workshop On Linked Spatiotemporal Data 2010: 14 September 2010, VENUE: Zurich
The event was attended by Kamie Kitmitto, the Landmap Service manager.
3. Free Open Data Master Class: 08 November 2010, VENUE: University of Newcastle
Gail Millin-Chalabi, the Geodata Research and Development Officer and Bharti Gupta attended this event and found the presentation from Talis very interesting and useful, which explained what Linked Data is about and how to work with data on the web.
4. Linked Data Hackdays: 13 February, 2011 -14 February, 2011, VENUE: University of Bristol
Bharti Gupta attended this two day event hacking content, systems and services for linked data and gained hands on training on RDF rapid development techniques.
5. DevCSI Developer Days: 16 February, 2011 – 17 February 2011, VENUE: London.
Various sessions dedicated on Linked Data were attended by Bharti Gupta to gain knowledge on Linked Data, also the idea was developed how Landmap metadata can be transformed into RDF.
6. Semantic Technologies For A More Powerful Web: 7 March, 2011, VENUE: Birmingham City University
Through this training from Netskills, James Schumm, the Geodata Services Software Engineer and Bharti Gupta explored the benefits of a more semantic web and how RDFa can be used to enhance the discoverability, usability and popularity of our online resources.
7. GeoVoCampSouthampton2011: 22 March, 2011 – 23 March, 2011, VENUE:Ordnance Survey
Bharti Gupta grouped with Peter Parslow from Ordnance Survey to discuss the metadata vocab for geo-spatial data. Emphasis was on ISO 19115/19139 standard metadata.
8. Linked Data with GI – Workshop and Tutorial, GISRUK 2011: 26 April, 2011, VENUE: University of Portsmouth
This one day workshop was attended by Yin Tun, the Geo-Metadata Development Officer, Gail Millin-Chalabi, Bharti Gupta and Kamie Kitmitto. It was good to learn the potential advantages, and see worked through Examples. We also shared our knowledge gained so far and learnt from other delegates’s work.
Progress made so far…
As our team developed a background about the Linked Data concept, some work was done using Java 6, JAXP and XSL to produce a sample RDF. Work is in progress to improve the XSL and work on a better source metadata XML which is conforming to ISO 19115/19139 standard.
ISO Standards 19115 for metadata and 19139 for XML implementation are being thoroughly studied not to leave any of the mandatory and other useful meta-tags. Knowledge gained from the GeoVoCamp is further enhanced and more vocabularies and ontologies are being used to develop the RDF.
Landmap plans to get the Project Workflow established by discussions and in-house workshops very soon. Specifications will be prepared, which will result in enhancing the conceptual bubble graph for the Landmap metadata and thus resulting in better and well formed RDF ready to be put in database.