Uk discovery-jisc-project-showcase

Project Showcase

Twitter hashtag: #ukdiscovery Wifi: WellcomeNet

AIM25: Open Metadata Pathway• AIM25: aggregating and cross-searching descriptions of

archives held in London. 16,000 descriptions on every key theme; subject, personal & corporate name and place indexing; 1.3m hits per month

• King’s College London and ULCC lead; 123 institutional partners

• OMP: JISC-funded, Feb-July 2011. Add 1,100 descriptions from 5 new partners including National Maritime Museum and Kew Gardens; run Linked Data across the records to improve user searching. Blog & Twitter updates: #Aim25news

• Focus groups and statistical analysis to test outcomes and inform forward action

AIM25 OMP: Progress so far

• Gate and Open Calais used to interrogate samples of records including their indexing. Pros and cons of each explored. OC selected

• Articulating the link between ISAD(G), Linked Data, Open Metadata and faceted searching

• Aim to test the value of Linked Data – does it improve searching; does it speed up data input and editing; does it enhance or supplant indexing?

• Questions include: how to deal with subject terms (use UKAT – out of date?); which URI scheme to use? How will the user benefit – what are the tangible, demonstrable gains?

http://openmetadatapathway.blogspot.com/

COMET – Cambridge Open MetadataWhat?

Releasing large subset of UL records under a Public Domain Data License

Identifying IPR history of our bibliographic data

Documenting process and releasing tools for others to do the same

Converting to useful linked RDF

Some as Marc21

Establishing a library triple-store

Link to OCLC ‘next-gen’ authority services

Why?

See what developers can do with our stuff

Move libraries forward with open data licensing

Gain in-house understanding of semantic web

Better realise value in records through contribution to the public domain

COMET – Cambridge Open Metadata

Where ?http://cul-comet.blogspot.com/ - project blog

http://data.lib.cam.ac.uk – datasets and triplestore (eventually)http://lib.cam.ac.uk/api - library API portal

Who?

Ed Chamberlain – [email protected] - @edchamberlain

How and when ?

1. License negotiation with Record vendors – 75% complete

2. RDF Conversion – 70 % complete

3. Triplestore and linked data web infrastructure - 60% complete

4. Documentation and supporting work - 30% complete

JISC

Contextual Wrappers http://context.collectionstrustblogs.org.uk/

packaging collections information as part of open metadata provision

Contextual Wrappersdeveloping collection level descriptions for aggregation

on the Culture Grid as an aid to resource discovery

related item recordsrelated item records

related publicationsrelated publications

related collectionsrelated collections

collection level description

17th centuryDutch art inmuseum a

collection level description

17th centuryDutch art inmuseum a

collection level description

17th centuryDutch art inmuseum b

collection level description

17th centuryDutch art inmuseum b

collection level description

17th centuryDutch materialin archive a

collection level description

17th centuryDutch materialin archive a

17th centuryDutch art

subject area inundergraduatestudy

17th centuryDutch art

subject area inundergraduatestudy

improved contextual information about collection scope and strength relevant to area of study

for further information, contact: David Scruton [email protected]

Jerome                    1/2

University of Lincoln

What's the point of Jerome?• Harvesting data from several existing library systems• Unifying and standardising bibliographic data• Supplementing /enhancing with existing Open Data• Releasing 200,000+ bib records under CC0• Immensely fast search using NoSQL d/b and Sphinx• Personalised portal• Awesome APIs!• Range of data formats: 

XML, JSON, MARC, RIS

http://jerome.library.lincoln.ac.uk/

Jerome                    2/2

University of Lincoln

What have we done already?• Now harvesting live MARC, Repository, e-journals data• Live search portal available: http://lncn.eu/jerome• Basic APIs created

What's still to come?• Records with explicit Open Data licence• Fully-documentated APIs• "Radical Personalisation" (machine 

intelligence, mixing desk, is it raining?)• Mashup with the COMET project

http://jerome.library.lincoln.ac.uk/

OpenART : Open metadata for Art Research at the Tate

OpenART has at it’s heart a dataset ‘The London Art World 1660-1735’, the result of several years research into People, Sales, Places and Sources from the period.

OpenART is a collaboration between the University of York, Tate and Acuity Unlimited.

OpenART will model the entities, likely using an event-based approach, store in a semantic-friendly way and expose the data for aggregation.

OpenART will explore data normalisation with vocabularies and ontologies, and propose methods for Tate and York to expose or encode their own data.

What now, when next …

Analysis and working with our researcher to understand the complex dataset so that we can produce a data model; implementing a database to help manage the input and data structureFinalising the data model, choosing a modeling approach and ontology/ies and blogging about it; validating with stakeholder use cases; considering entity registration.

Contact: Julie Allinson [email protected]://tinyurl.com/dlib-openart -- http://yorkdl.wordpress.com/

Existing process

…Institutional Resolvers

OpenURL Router

Institutional Resolvers

Institutional Resolvers

Institutional Resolvers

Request

Redirect request

Logrequest

Level 0 Data(Log)

Survey institutions to enable opt-out

Level 1 Data

Process to Level 1:• Exclude data from opted-

out institutions• Anonymise IP addresses• Anonymise institution &

remove button & lookup data that would identify institution

• Parse OpenURL request into constituent parts

Process to Level 2:• Include only redirect to

resolver requests

Level 2 Data

Use for prototypes & services

Using OpenURL Activity Data: Overview

• Build on existing process for OpenURL Router log• Legal investigation: risk management and data

disclosure when processing IP addresses

Project aims:•Make this data

available under open licence

•Develop prototype service using this activity data

•Explore including institutions’ data in aggregation

OpenURL Activity Data: Project and Info

•What’s in the data set?– Date & time– Anonymised IP address &

institution IDs– OpenURL request data, e.g.

• Article Title• Journal Title• Book Title

•How might the data be used?– Article/journal recommendations– Student analysis– Research thesis– Publishers comparing listings with

texts sought– Innovative services to meet your

users’ needs– Other, unanticipated uses

Legal investigation: report available

Licence selection: ODC-PDDL with Attribution

Sharealike community norms

Extend openurl.ac.uk for provision of data

Aggregated Level 2 Data

Available

Key activity to date: Next steps/outputs:

Provide Level 1 & Annual Data

We are here:

Develop Prototype Recommender

Service

• Author• ISSN• DOI …

•Further information:– openurl.ac.uk/doc– EDINA Project Page: Using OpenURL Activity Data– Sheila Fraser: [email protected] 0131 651 7715

SALDA ProjectSussex Archive Linked Data Application

http://blogs.sussex.ac.uk/salda/

Transforming the catalogue records of the

Mass Observation Archive into Linked

Open Data

SALDA ProjectProgress

• Converted catalogue records to EAD • Ongoing transformation of data to Linked Data

by Pete Johnston at Eduserv • Using a PDDL licence to make data available• Using data.lib.sussex.ac.uk as our URI stemMore information at:

http://blogs.sussex.ac.uk/salda/Contact: Karen Watson [email protected]