Semantic Web underpinnings of the IRI Data Library

1
Semantic Web underpinnings of the IRI Data Library

• Semantic Web as a Framework for Multiple Metadata
• IRI Data Library presenting Data in multiple
  frameworks
• Pure RDF Faceted Data Search
• Metadata interoperability using adiabatic XML
  Schema to OWL mapping

http//iridl.ldeo.columbia.edu/ontologies/
2
blind monks examining an elephant
Multiple partial representations of objects
described by data
John Godfrey Saxe (1816-1887)
3
Semantic Web as a Framework for Multiple
Metadata The Data Problem
Datasets
Users
4
The Tool Interface
Datasets
Tools
Users
5
Standard Metadata
Standard Metadata Schema/Data Services
Datasets
Tools
Users
6
(No Transcript)
7
Super Schema
Standard metadata schema
8
(No Transcript)
9
(No Transcript)
10
Where we are coming from the IRI Data Library

• presenting Data in multiple frameworks

11
(No Transcript)
12
(No Transcript)
13
(No Transcript)
14
(No Transcript)
15
(No Transcript)
16
IRI General Data Tools
Data Page
17
IRI General Data Tools
Data Viewer
18
Calculations svd
(link svdview)
(link svd results dataset)
(link svd documentation)
19
WMS and KML land cover
(linkfigure page)
20
Malaria Early Warning System
IRI Map Room
Front page illustrates most recent dekadal
rainfall estimates (FEWS RFE)
Administrative and epidemiological overlays
available
Change dates to view different time periods
Click and drag box across map to zoom
21
IRI Data Library Faceted Search an example of

• RDF-based faceted search for climate data
• Drawing on multiple ontologies to build an
  application
• Using inference to connect ontologies describing
  different parts of the framework

22
Search Interface as Multiple Ontologies
Additional Semantics
Dataset Ontology
Search Ontology
Datasets
Search Interface
Users
23
Faceted Search
http//iridl.ldeo.columbia.edu/ontologies/query2.p
l?...
24
Distinctive Features of the search

• Search terms are interrelated
• terms that describe the set of returns are
  displayed (spanning and not)
• Returned items also have structure (sub-items and
  superseded items are not shown)

25
Architectural Features of the search
http//iridl.ldeo.columbia.edu/ontologies/query2.p
l

• Interface is generated from a set of Terms
  connected to a set of Items
• Multiple search structures possible
• Multiple languages possible
• Search structure is kept in the database, not in
  the code

26
Term Ontology

• Concepts as individuals (unlike conceptual
  ontology with concepts as classes)
• Simple Knowledge Organization System (SKOS) is a
  prime example
• The ontology used here is slightly different than
  SKOS facets are classes of terms rather than
  being top_concepts

27
Nuanced tagging

• Concepts as objects can be interrelated specific
  terms imply broader terms
• Object ends up being tagging with terms ranging
  from general to specific.
• Search can then be nuanced
• tagging can proceed in absence of perfect
  information

28
Faceted Search Explicated
29
Search Interface

• Items (datasets/maps)
• Terms
• Facets
• Taxa

30
Search Interface Semantic API

• item dctitle dcdescription rsslink
  iridlicon
• dctermisPartOf item2
• dctermisReplacedBy item2
• item trmisDescribedBy term
• term a facet of taxa of trmTerm,
• facet a trmFacet, taxa a trmTaxa,
• term trmdirectlyImplies term2

31
Faceted Search w/Queries
http//iridl.ldeo.columbia.edu/ontologies/query2.p
l?...
32
(No Transcript)
33
IRI RDF Architecture
Data Servers
MMI
Ontologies
JPL
Start Point
bibliography
Standards Organizations
RDF/XML-Schema Crawler
Location Canonicalizer
XML Schema to OWL translation Owl Semantics SWRL
Rules SeRQL CONSTRUCT
Time Canonicalizer
Sesame
Search Queries
Search Interface
34
Cast of Characters

• NC netcdf data file format
• CF Climate and Forecast metadata convention for
  netcdf (includes controlled vocabulary for
  physical quantities)
• SWEET - Semantic Web for Earth and Environmental
  Terminology (OWL Ontology)
• IRIDL IRI Data Library

35
CF attributes
NC basic attributes
IRIDL attributes/objects
CF data objects
CF Standard Names (RDF object)
SWEET Ontologies (OWL)
Location
IRIDL Terms
CF Standard Names As Terms
SWEET as Terms
Search Terms
Gazetteer Terms
36
XML Schema to OWL ontologies

• Semantically-mediated translation between Web
  Coverage Service (WCS) and OPeNDAP (data as
  multiple-dimension variable)
• WCS is defined in XML Schema (57!)
• Sample Schema to OWL Translation
• Method to return JDOM element (XML) containing
  requested XML property from RDF triple-store
  (element maps to (Resource,class) pair)
• Sample Datasets/Coverages

37
Please Visit

• IRI Data Library http//iridl.ldeo.columbia.edu/
• IRI ontology work http//iridl.ldeo.columbia.edu
  /ontologies/
• Or contact
• Benno Blumenthal
• John del Corral
• Haibo Liu