Scraping Toward A Giant EduGraph
Patrick Murray-JohnACCS 2009Charlottesville, VAMarch 11, 2009
Semantic UMW
A Sad Story . . .
CHEMISTRY 101
CHEMISTRY 101
CHEMISTRY 101ENGLISH 205
YOUR DATA!! SHALL NOT!! PASS!!
CHEMISTRY101ENGLISH205
Here's the part that makes me cry
Flickr image by Addrox CC-BY-NC-ND
BOOKSTORE ORDER:CHEMISTRY 101
BOOKSTORE ORDER:ENGLISH 205
UNIVERSITY
BOOKSTORE
BOOKSTORE ORDER:CHEMISTRY 101
BOOKSTORE ORDER:ENGLISH 205
UNIVERSITY
BOOKSTORE
UNIVERSITY
BOOKSTORE
HELP ME!
Here's the part that makes me cry AGAIN
SYLLABUS:CHEMISTRY 101
SYLLABUS:ENGLISH 205
DEPARTMENT
CHAIR
Why am I crying?
Flickr photo by ifindkarma CC-BY
Flickr photo by manu contreras CC-BY
Instead. . .
Focus on the data, not the document, and links can go anywhere
CHEMISTRY 101Leana daProf
Alice daStudent
Chem101Course Blog
Thankyou!
A Giant EduGraph
Giant Global Graph
Semantic Web (aka Giant Global Graph)
Linked Open Data
Aggregating / ExposingHigher Education Info
CoursesCourse GroupsInstitutions / DivisionsOnline SpacesPlacesEvents / LecturesLibrary / Book DataClubs / OrganizationsRequirements / Credentials
Flickr photo by smannion CC-BY-NC-SA
Image from Wikipedia GNU-FDL
Semantic Web
Identify things with a URI (a node in a graph)
Frankenstein(URI)
Semantic Web
Data is Nodes + Relationships between them
CHEMISTRY101 S09-Group(URI)Frankenstein(URI)Alice daStudent(URI) Studies textMember
Linked Open Data
Semantic Web + some technical goodiesOpen on web
Linked Open Data
Flickr image by Rgis Gaidot CC-BY-NC
What Will It Look Like?
Taps into LOD Cloud
DBpedia references for topics, peopleevents, etc.GeoNames for places, buildings, etc.OpenLibrary for book infoDomain-specific sources (MusicBrainz,Linked Movie Database, more)
What Will It Look Like?
Uses existing vocabulariesSIOC - Semantically Interlinked OnlineCommunities FOAF - Friend Of A FriendDCTERMS - Dublin CoreBIBO - Bibliographic InfoAIISO - Academic Institution InternalStructure Ontology. . . and more, plus some being developed
Where Will The Data Come From?
Scrape existingstructured data
Courses andCourse Groups
Places
People
Requirements & Credentials
Where Will The Data Come From?
Zotero
(actually Bruce D'Arcus' )
likely to expose SW-friendly data in the future
Where Will The Data Come From?
-- 13 years worth of circulation data anonymized and published as XML
-- took about 3 days to RDFize it and make an Exhibit
Follow lead of Univ. of Huddersfield
Where Will The Data Come From?
Talis Project Xiphos
-- Putting UK university reading lists into RDF
Where Will The Data Come From?
Need an easy interface to relate course group data to LOD data (lookup.dbpedia.org)
. . . and to relate people and course groups with online spaces
. . . and to add data that isn't otherwise available
Biggest Challenge
Biggest Challenge
OPENNESS
Top Related