Long journey of Ruby standard library at RubyConf AU 2024
Experience from 10 months of University Linked Data
1. Experience from 10 months of University Linked Data Mathieu d’Aquin - @mdaquin Knowledge Media Institute, the Open University LUCERO project lucero-project.info – data.open.ac.uk
2. Linked Data As set of principles and technologies for a Web of Data Putting the “raw” data online in a standard, web enabled representation (RDF) Make the data Web addressable (URIs) Link with other data
4. The Open University The biggest university in the UK (200,000 students) One of the youngest (40 years) Most teaching done at a distance 1 campus, 13 regional centers Committed to “Open”: Open educational material available as podcasts (iTunes U), units of course material (OpenLearn), etc. Tradition of investing in new technology for teaching, learning, knowledge sharing, etc. Role of the Knowledge Media Institute (KMi)
5. So Linked Data for the OU? RAE DBPedia Data from Research Outputs OpenLearn Content ORO Exposed as linked data, our data interlink with each other and the external world: become part of the “global data space” on the Web Archive of Course Material Library’s Catalogue Of Digital Content geonames data.gov.uk Currently: OU public data sit in different systems – hard to discover, obtain, integrate by users. A/V Material Podcasts iTunesU BBC DBLP
6. Why is it important? The OU has been the first University to expose its data as linked data: http://data.open.ac.uk Now widely recognized as a critical step forward for the HE sector in the UK (and worldwide) Favor transparency and reuse of data, both externally and internally Reduces cost of dealing with our own public data: integration and reuse by design Enable both new kinds of applications, and to make the ones that are already feasible more cost effective At least 3 other UK universities have now followed our example: http://data.online.lincoln.ac.uk/, http://data.ox.ac.uk/, http://data.southampton.ac.uk/ And others in other countries are setting up similar initiatives
7. The data.open.ac.uk Stack Applications Institutional repository data Research Data (Arts) Organizational infrastructure Technical infrastructure
9. Expose Store Collect Extract Link Ontologies Scheduler Cleaning rules RDF file (add) RDF file (delete) URL redirection rules RSS Extractor Delete (1) Add (2) RDF Cleaner Web Server ORO, podcast RSS feed RDF file (add) RDF file (delete) Triple Store RSS Updater SPARQL endpoint RDF Extractor New items Obsolete items Each datasets Index Entity Name System Search XML Updater URI creation rules Lib, courses, loc Planning + Logging Generic process Dataset specific process
17. Define URI SchemeData Modeling Validation Lucero Core Team Lucero members Data Owner Development of Extractor URI Creation Rules Definition Deployment Lucero KMi Team
20. Applications For education Mobile podcast explorer, podcast explorer on TV OU Building Map, OU location tracker (cf. foursquare) OU Expert Search Connecting courses/OpenLearn to relevant podcast OU Course Profile Facebook app using list of courses, “Study Buddy” app connecting facebook users to relevant courses For Research Display connections in a research community Research Data/Impact Analysis Connection research datasets to external data
24. Example application: Expert Search using publication information and connecting to contact information within the OU
25. Example application: Explore Information about a person in the “Reading Experience Database” based on data provided by DBPedia (Linked Data version of Wikipedia) New ways to look at humanities research data
26. Lessons Learnt The major part of the work is not technical Linked data is simple! Identifying available data, obtaining access to them, re-modeling them is hard Making people understand that it is worth doing is critical Especially when dealing with challenges such as data licenses, private data, etc. Get people involved (it is not about you, or the technology) A lot of people’s job (administrators, managers, researchers) is all about collecting and managing data A lot of this effort is lost because of closed systems, lack of integration and exposure of the data Our job is to demonstrate to these people how the principles of linked data can be used to leverage this effort Without being disruptive (e.g., the URI of a course in a browser redirects to the course webpage on the OU website
27. Lessons Learnt There is no killer app The direct benefit of linked data is not in a great big smart application, it is in the many small things that are made easier Need to make it easy for developers to get into it, play with it, see the potential by themselves Integrating the benefits of linked data in the university’s practices/workflows takes time. It is not a threatening big change, but a slow, incremental adoption Plan for long term = need for endorsement We work with the assumption that, soon, it will be as common and necessary for a University to have a linked data platform as it is to have a website So a linked data initiative at a university cannot be a one time thing. Courses evolve, new material appear, new datasets are made available. (e.g., data.open.ac.uk is updated every day) It needs to become part of the University’s role and be endorsed by the departments involved (IT, communication, education, research, business) It does not always work Some applications might be incompatible with the University’s policies (e.g., Google rich snippet showing the price of a course) Support might only get up to a certain point
28. The future From nice demonstrators to real semantic web applications Use of reasoning and data mining for data consolidation and analysis Need proper frameworks for application developers! Linked data and the Semantic Web to support research Not only research communities Identifying new research questions and collecting evidence through connected datasets It is not about individual Universities! Universities sharing data to benefit students and researchers: the higher education’s web of linked data Needs collective vocabularies, recipes, approaches, classifications… the GoodRelations of higher education?
29. The future Linked data analytics/Linked data mining Interfaces to linked data/Making sense of linked data (with ontologies) Semantic web for activity data/personal data
30. Thank you! Carlo Allocca (Dev) SalmanElahi ((Ex)-Dev) Jane Whild (Admin) FouadZablith (Dev) KMi AndriyNikolov (linking) Enrico Motta (SGP) Mathieu d’Aquin (PD) Arts Suzanne Duncanson-Hunter John Wolfe Paul Lawrence Richard Nurse ((ex-)PM) Owen Stephens (PM) Stuart Brown Com./ Student Comp. Services Data Owners Non Scantlebury Library Specialists Arts Specialists OU Library
Notes de l'éditeur
Usual pitch: - data on the web = every piece of data is web addressable, so data across different places/stores/systems become linkable: the Web = 1 data space