LUCERO - Building the Open University's Web of Linked Data


Published on

Seminar given at KMi, the Open University about the LUCERO project ( and the process of building
Video available at

Published in: Technology
  • Be the first to comment

  • Be the first to like this

LUCERO - Building the Open University's Web of Linked Data

  1. 1. Building the Open University’s Web of Linked Data Mathieu d’Aquin
  2. 2. Linked Data • As set of principles and technologies for a Web of Data – Putting the “raw” data online in a standard representation (RDF) – Make the data Web addressable (URIs) – Link with to other Data
  3. 3. Linked Data
  4. 4. Linked Data at the OU? ORO Archive of Course Material Library’s Catalogue Of Digital Content OpenLearn Content A/V Material Podcasts iTunesU Data from Research Outputs BBC DBPedia DBLP RAE geonames
  5. 5. Goals of the project (visually) Technical infrastructure Organizationalinfrastructure Institutional repository data Research Data (Arts) Applications
  6. 6. Institutional Datasets • Open Research Online (ORO) • Library catalogue • Study at the OU • iTunesU/OU Podcasts • Staff profiles • OpenLearn • Estate Information
  7. 7. Research datasets • Looking at how specific research databases can benefit from being linked to the institutional repositories • Case studies in Arts: – Classical Receptions in Drama and Poetry in English – Open Arts Archive – Encyclopedia of Global Commodities – Hestia – Reading Experience Database – The South-Asians Making Britain project Research Data OU linked data cloud Web of data
  8. 8. Planning + Logging Collect Extract Link Store Expose OntologiesScheduler RSS Updater Delete (1) Add (2) Search SPARQL endpoint Web Server RSS Extractor XML Updater RDF Extractor RDF Cleaner Cleaning rules Each datasets Lib, courses, loc ORO, podcast URL redirection rules RSS feed New items Obsolete items RDF file (add) RDF file (delete) RDF file (add) RDF file (delete) Generic process Dataset specific process Entity Name System URI creation rules
  9. 9. First Version of with 2 datasets: ORO and Podcasts
  10. 10. Dataset: ORO • Open Research Online • Scientific publications with at least one member of the Open University as co- author • Original System based on ePrints • Export to RDF using the BiBO Ontology • Post-processing/cleaning • 13,283 Articles/12 Patents/340,000 triples
  11. 11. Dataset: Podcast • Extracted from RSS feeds at • Using W3C Media Ontology, FOAF, DCT, Media RDF, etc. • Provides connections to courses and topics • 1,664 Video Podcasts/1,325 Audio Podcasts/75,000 triples
  12. 12. Workflow for a given dataset Initial Meeting with Data Owner - Identify data - Get sample data - Identify Copyright Issues - Identify possible links - Identify users and usage Data Modeling sessions Lucero Core Team Data Owner Lucero KMi Team Lucero members - Find reusable ontologies - Map onto the data - Identify uncovered parts - Define URI Scheme Data Modeling Validation Lucero Core Team Data Owner Development of Extractor URI Creation Rules Definition Deployment Lucero KMi Team
  13. 13. Ongoing Dataset: Courses<saou_course_data> <Course> type: courseware:Course, aiiso:Module <CourseCode>D240</CourseCode> URI of the object : (+ in property aiiso:code) <CourseTitle>Counselling: exploring fear and sadness</CourseTitle> // courseware:has-title, aiiso:name, rdfs:label, dct:title <OUCourseLevel>2</OUCourseLevel> // new property oco:OUCourseLevel (domain aiiso:Module, range: int). <CourseUGPG>Undergraduate</CourseUGPG> // new property oco:courseLevel (domain aiiso:Module, range: CourseLevel). // New Class CourseLevel, 2 instances || <CourseStatus>L</CourseStatus> // courseware:is-taught-present = true // will need additional checks for making courses unlive... <ScottishLevel>8</ScottishLevel> // new property: oco:SCQFLevel (d aiiso:module r int) <CreditPoints>30</CreditPoints> // oco:number-of-credits subpropertyof courseware:number-of-credits <ECTSCreditPoints>15</ECTSCreditPoints> // oco:eu-number-of-credits subpropertyof courseware:number-of-credits // (label for ey-number-of-credits = "Number of credits according to the ECTS) <QAACode>5</QAACode> // new property: oco:FHEQLevel (d aiiso:module r int) <IsCPD>N</IsCPD> // if N, do nothing, if Y than type of course is oco:CPDModule (subclassof aiiso:Module and courseware:Course) <IsOnline>N</IsOnline> // if N do nothing, if Y type is oco:OnlineCourse (subclass of aiiso:Module and courseware:Course) and // if code.contains("ZX"), oco:isOnlineVersionOf=http://data.../course/<code.replaceAll("ZX","")> <FacultyUnitCode>FD15</FacultyUnitCode> // URI: = new instance of aiiso:Faculty <FacultyUnitName>Social Sciences</FacultyUnitName> // rdfs:label for the faculty // course to faculty linked by + courseware:involves-organization // + course courseware#taught-at <ResidentialSchool>N</ResidentialSchool> // if Y: oco:hasAssociatedEventType // (type: oco:EducationalEventType) <ResidentialSchoolType>No residential school</ResidentialSchoolType> // if No residential school = nothing // if Embedded residential school oco:hasAssociatedEventType // (which is oco:subEventType residentialSchool) // Stand-alone residential school oco:hasAssociatedEventType // (which is oco:subEventType residentialSchool) <NumberOfCMAs /> // if 0 or nothing do nothing, else courseware:has-assesment-method // (which is type courseware:Assessment-Method) <Exam>End-of-course assessment</Exam> // if No examination do nothing // if End-of-course assessment courseware:has-assesment-method // (which is type courseware:Assessment-Method) // if Examination courseware:has-assesment-method (which is type courseware:Assessment-Method) <ECMAsTotal /> // if 0 or nothing do nothing, else courseware:has-assesment-method
  14. 14. Course RDF representation: Basics …
  15. 15. Course RDF representation: Locations …
  16. 16. Course RDF representation: Prices
  17. 17.
  18. 18. Applications • Plan for development of specific applications targeting: – Students: in finding resources related to courses, topics, and helping selecting courses to enroll to – Researchers: Identify interesting connections/research questions from research data linked to OU/external sources • Already a number of (more generic) applications emerging…
  19. 19. Faceted Browsing
  20. 20. Social Learn • … SocialLearn (Recommendation)
  21. 21. Next • More data, more data, more data… – Info about buildings (estate), course material (Library), open educational content (OpenLearn), People (AACS/IT), Anything else? • Links, Links, Links… – Internal links based on common URI Scheme – Current experiments on linking people, conferences, journals – Automatic discovery of links • Applications, Applications, Applications – A killer app (SocialLearn? Research Data?) – Many, many small usages (publication and courses in people profiles, “related material” in course pages, annotation, etc.) – Anything KMi can come up with!?
  22. 22. Conclusion “[] is the first launched of its kind in UK Universities” “hopefully this is the first of many to come!” “if you are working in an IT department within a University you better read this report, as soon your department will need to be making these same decisions.” David Flanders, JISCExpo Programme Manager,
  23. 23. Conclusion • We believe linked open data to be essential to the OU, so it needs support: – To ensure a sustainable framework to expose, maintain, update and evolve Open Linked University Data – To adopt the practices of exposing data a common, reusable, open way through – To develop the applications that will show the benefit of the technology
  24. 24. Thanks! Carlo Allocca (Dev) Mathieu d’Aquin (PD) Salman Elahi ((Ex)-Dev) Enrico Motta (SGP) Andriy Nikolov (linking) Jane Whild (Admin) Fouad Zablith (Dev) Library Specialists Owen Stephens (PM) Richard Nurse ((ex-)PM) Non Scantlebury Arts Specialists Suzanne Duncanson-Hunter John Wolffe Paul Lawrence Stuart Brown Data Owners KMi OU Library Com./ Student Comp. Services Arts