The Winner Takes it All? -APIs and Linked Data Battle It Out
Download as PPTX, PDF4 likes1,775 views
This document summarizes a presentation about aggregating and linking data from various sources about World War 1. It discusses creating APIs to expose data from different institutions, using formats like SOLR, RSS, and OAI-PMH. It also describes efforts to link related data across sources using Linked Data principles and forming a discovery layer to make the aggregated content more discoverable to people and machines. Challenges discussed include varying data formats, lack of certain content types, and non-open licenses.
The Winner Takes it All? -APIs and Linked Data Battle It Out
- 1. The Winner Takes It All? APIs and Linked Data Adrian Stevenson Jane Stevenson Mimas, The University of Manchester
- 3. WW1 Discovery Project • Exemplar illustrating principles of the JISC Discovery initiative • Discovery about advocating ‘open’ and ‘aggregating’ • Make digital content more discoverable by people and machines • Building WW1 aggregation API www.discovery.ac.uk and discovery layer
- 4. WW1 Discovery: How? • Aggregate data from existing APIs – IWM and NMM • Help others with example APIs – BL, MCR Archives, Welsh Voices, LSE • Formats: SOLR, RSS, OpenSearch, OAI- PMH, CSV
- 5. Prototype WW1 API #sxsw #lodlam
- 6. WW1 Discovery Layer: Info
- 7. WW1 Discovery Layer: Geo-data
- 8. Some Challenges • Difficulties merging data – varied content • Lack of content – images – geo-data • Content licenses not open • Lack of APIs
- 9. Linking Lives • Linking Lives is a JISC-funded project to create an end-user interface based on Linked Data • A biographical interface, providing information about individuals that is taken from a variety of sources • It enables us to place archival descriptions within a much broader context
- 16. Event: Birth of Skinner, Beverley, 1938-1999, artist and Death of Skinner, Beverley, 1938-1999, artist
- 17. Martha Beatrice Webb Image Works Life dates: 1858-1943 Epithet: social reformer and Our Partnership historian My Apprenticeship Family name: Webb The case for the factory acts Beatrice Webb’s diaries; edited by Margaret Cole The Diary Place of birth: Gloucester, England Place of death: Liphook, Knows Hampshire, England Biographical Notes from: Beatrice Webb letters Beatrice Webb (1858 - 1943). Fabian Socialist, social reformer, writer, historian, diarist. Wife, collaborator and assistant of Sidney Webb, later Lord Passfield. Together they contributed to the radical ideology first of the Liberal Party and later of the Labour Party. http://dbpedia.org/page/George_Bernard_Shaw from: Beatrice Webb, A summer holiday in Scotland, 1884. Beatrice Webb (1858-1943), nee Potter, social reformer and diarist. Married to Sidney Webb, pioneers of social science. She was involved in many spheres of political and social activity including the Labour Party, Fabianism, social observation, investigations into poverty, development of socialism, the foundation of the National Health Service and post war welfare state, the London School of http://dbpedia.org/page/Sidney_Webb,_1st_Bar on_Passfield
- 20. Grace Ford, 1896-1941 ‘same as’ Grace Ford? http://archiveshub.ac.uk/data/gb738ms58 http://viaf.org/viaf/44311096/
- 21. Hub data inconsistencies • Winston Leonard Churchill • Sir Winston Leonard Spencer Churchill • Churchill, Sir, Winston Leonard Spencer, 1874- 1965, knight, prime minister and historian • Churchill, Winston Leonard, 1874-1965, prime minister • Churchill, Sir Winston, 1874-1965, knight, statesman and historian
- 22. We can start to say things that go beyond what is known within our own space…and if we use the same URIs we can link data sources much more easily http://data.archiveshub.ac.uk/id/person/nra/webb marthabeatrice1858-1943socialreformer <is the same as> http://viaf.org/viaf/86607236/ <is the same as> http://bnb.data.bl.uk/doc/person/WebbBeatrice1858- 1943
- 23. Vocabularies
- 26. Conclusions • APIs offer useful tools & lightweight approach • WW1: creating APIs for institutions • Locah: Linked Data for over 200 institutions • Linked Data makes use of the Web architecture • Uses HTTP URIs to represent resources • Navigate through things via URIs – Web of Data • Can make APIs more ‘Webby’
- 27. Conclusions • Linked Data opens data up to wide variety of uses • Can de-reference classes and properties • Need to think more about the end user • Need more tools • Need to collaborate
- 31. Maybe not so much…. http://www.funnyjunksite.com/pictures/funnypics/sport/sports/funny_s ports_picture_99.jpg
Editor's Notes
- #2: A presentation given at EmTACL 2012 in Trondheim.
- #3: WW1 Discovery WebsiteThis is home for our WW1 API project. Specifically looked at lightweight APIs.Say that an API stands forKind of back for machines to get access websites and services’ content
- #4: About the WW1 Discovery project
- #5: Using APIs and helping institutions set up APIs.
- #9: Challenges include the lack of APIs available to aggregate data.
- #10: Linking Lives takes a different approach – using Linked Data rather than merging data through APIs.
- #11: The core data comes from the Archives Hub, UK aggregator of archival descriptions
- #12: The Archives Hub provides classic search/retrieve of ISAD(G) archival finding aids
- #13: Descriptions are often rich in content – often extensive scope & content and biographical history. Controlled index terms – subjects and names. Every descriptions has an HTTP URI.
- #14: Often multi-level descriptions.
- #15: The Locah project transformed Archives Hub EAD (XML) descriptions into RDF. The homepage provides access to an RDF dump, a Sparql endpoint and the XSLT stylesheet.
- #16: Entities such as repositories have a number of triples describing relationships, including spatial data.
- #17: Triples relating to individuals. Include modelling of birth and death dates as events.
- #18: Mock-up of the LInking Lives interface shows the way data is brought together.
- #19: External data is key to linked data. We link to VIAF and through that to DBPedia. We are looking at linking to the BNB.
- #20: But matching strings is not easy, e.g. matching subjects in the Hub with subjects in LCSH.
- #21: Our solution for matching names in the hub to names in VIAF was quite manual – it cannot be used when we up-scale. Many names are hard to match – may not have life dates.
- #22: Names are often entered into the Hub in different ways, despite the use of Rules.
- #23: The power of linked data is the way we can match up data from different datasets. We could consider using the same URIs, e.g. VIAF URIs for names.
- #24: One of the challenges of doing LInked Data is the plethora of vocabularies. It is hard to decide what we should use.
- #25: Shared RDF model could really help – still deal with different domain models (ontologies) – eg. BL data and Locah data are quite close but different models. So not just a ‘perfect fit’ – and others may be modelled more differently. Common ‘meta model’. Otherwise pulling in metadat from umpteen different metadata formats (cos not all using RDF). Eg. pull in TEI, MODS, EAD, MARC data.
- #26: Something like the BL model could be used as the basis for other bibliographic data.
- #27: – the way websites work is the way Web APIs should work, but not always the case – doesn’t have links in. Often you have to know how to create the links. They don’t use the power of the Web as a whole.API often chunk of XML with numbers in and you have to know what to do with themThere are substantial benefits to participating in the existing network of URIs, including linking, bookmarking, caching, and indexing by search engines, and there are substantial costs to creating a new identification system that has the same properties as URIs.
- #28: Linked data offers a great deal but maybe not thinking enough about end-users yet.
- #29: the legislation.gov.uk site is a good example of a site using linked data effectively.
- #30: It provides effective searches…but these benefit from an API on top of the data…
- #31: So it is an exmaple of the benefits of utilising both linked data and APIs.
- #32: It is probably not constructive to think in terms of APIs vs. Linked Data
- #33: They should work together! x