ESDSWG2012 Semantic Tech. Impacts Semantic Technologies:
Activities and Impacts (TIWG Sub-Group) ESDSWG Meeting November 13
15, 2012 Annapolis, MD Brian Wilson (Mike Little) Hook Hua
Slide 2
Semantic Technologies 2 Semantics more than just OWL/RDF
semantic web stack Web mining using text understanding
(auto-classification) Topic & keyword extraction: DBpedia
Spotlight, OpenCalais, etc. Linked Open Data (LOD) RDF data graphs
published on open Internet Data integration by simply merging
triples SPARQL query endpoints, explore logical triples Return JSON
for web mashups Extending Ontologies (OWL) E.g.: Concepts from
SWEET, Noesis, W3C PROV for provenance Semantic mediation using
ontologies Query terms broaden/narrow, synonyms, OWL SameAs Graph
Visualization Provenance, collaboration, joint authorship, social
network
Slide 3
Semantic Web Stack 3 Linked Open Data (LOD)
Slide 4
Resource Description Framework (RDF) Built on the logical
triple, a 3-tuple consisting of Subject, Predicate, and Object
Example graph below: Resource: Some entity. Property: An attribute
of a resource. Literal: A string of characters which can be the
value of a property. 4. "August 16, 1999". "en".
Slide 5
TouchGraph: Facebook social graph 5
Slide 6
Activities of the Group 6 Developing killer apps for semantics
Demo session at Jan. 2012 ESIP Federation meeting Evolving best
practices and tools experience Jena API, Protg, COE, Virtuoso,
SPARQL, LODSpeakr Extending ontologies (joint with ESIP Semantic
Web Cluster) Future of SWEET maintenance OWL/RDF vocabularies for
the AGU app. Publishing NASA metadata as Linked Open Data (LOD)
Prototype publication of LOD for ECHO collections, GCMD services
New interfaces: Instant Browse of ECHO collections Proposed
(future) community task for consideration: Publishing NASA Metadata
for Semantic Mashups
Slide 7
Linked Data Publication (5 stars) 7 Make your stuff available
on the web (whatever format) Make it available as structured data
(e.g. excel instead of image scan of a table) Non-proprietary
format (e.g. csv instead of excel) Use URLs to identify things, so
that people can point at your stuff Link your data to other peoples
data to provide context LOD is about authoring rich linkages
between data Use of URIs/URLs to permanently identify resources is
crucial! Things that should have a permanent URL: Every NASA
dataset registered in GCMD & ECHO All registered services and
tools AGU authors, meeting abstracts, sessions AGU journal articles
(DOIs)
Slide 8
Linked Open Data 82011-01-06T10:15:00-05:00
Slide 9
Benefits of Open Metadata 9 Both LOD and casting technologies
free our metadata Openly published on the web in interoperable
formats Objects are permanently named by URI GCMD, ECHO, DAACs
should use their HTTP namespaces The real estate grab on the web
has already occurred Any third-party can make logical assertions
about your objects Crowdsource rich linkages between metadata
objects E.g.: tools to datasets, datasets to all relevant web
services, science papers to datasets used, granule to QC
annotations, people/projects graph to find new collaborators.
Semantic Mashups as killer apps Event Linkages: phenomenon, data,
services, tools, human impacts [Your APP here.]
Slide 10
Application Demonstrations 10 Demo session at Jan. 2012 ESIP
meeting Demo session at Jan. 2012 ESIP meeting Linked Open Data for
AGU Abstracts, Sessions, & People (Eric Rozell, Tom Narock)
LOD, text understanding (Dbpedia), SPARQL The ESIP Collaboration
Network (Erin Robinson) social graph, visualization, data
integration Noesis 2.0: Smart Search for Collections & Services
(Rahul Ramachandran) query expansion by ontology, meta-search,
topics, user taxonomies Saving & Querying Production Provenance
Graphs using the Earth Science extension to W3C PROV standard (Hook
Hua) PROV-ES ontology, RDF graphs, SPARQL, viz Graph Visualization
Tools (Ruth Duerr, Joe Glassy) RDF viz. tools
Slide 11
Application Demonstrations (2) 11 DQSS: Data Quality Screening
Service (Chris Lynnes) Ontology for applying pixel-level quality
screening Spy Glass - Ontology based text mining (Rahul
Ramachandran, John Rushing) Hybrid system for text
understanding
Slide 12
AGU Meeting at a Glance: Web app. for iPhone and browser AGU
App: Sessions, Abstract, Authors 12
Slide 13
OWL/RDF Vocabularies 13 foaf Friend of a Friend dc Dublin Core
swrc Semantic Web for Research Communities swc Semantic Web
Conference tw Tetherless World geo WGS84 Lat/Long sweet Semantic
Web for Earth & Environmental Terminology skos Simple Knowledge
Organization System xsd XML Schema ao Annotation Ontology
PROV-ES: Processing Graph 16 Each process Uses its input files
and Generates its output files ACCESS-2009 project, PI Hook Hua
(JPL)
Slide 17
App. Demos in Telecons 17 VisKo: Semantic Auto-Composition of
Visualization Workflows (D. Pennington, Nicholas Del Rio; UTEP)
Automate format conversions, selection of viz. type, data
transforms Instant Browse of ECHO Collection Metadata (B. Wilson)
Publish as LOD for mashups (prototype) ToolMatch (C. Lynnes) Match
datasets to tools that manipulate them Session at the July ESIP
Federation meeting
Recommendations 20 Openly Publish All NASA Metadata Collection
& Data Casting, Service Casting, Event Casting Linked Open Data
Continue to Develop Killer Mashups Tools to datasets Datasets to
all relevant services (OpenSearch, WMS, DAP) Science publications
to datasets/variables used Granules to Quality Control annotations
(by DAACs & users) AGU/ESIP People/papers/projects
collaboration graph recommend new collaborators and papers to read
[YOUR App HERE.]
Slide 21
Publish ECHO Collections as LOD 21 LODSpeakr and other tools
provide generic browse or custom apps.
Slide 22
Instant Browse of ECHO Collections 22 Drill-down by semantic
facets and keyword search, with instant results
Slide 23
Instant Browse of GCMD Services 23 Drill-down by semantic
facets and keyword search, with instant results
Slide 24
Carbon Cycle Cyclone Tracks for Western Pacific in 2011 with
A-Train L2 Variables Event Cast Browser with KML Layers
Slide 25
Carbon Cycle Event or Pattern Observed Publish Event Cast w/
geoloc & time Subscribers Receive Announcement Event Cast
format: Atom, xlink Standards & Tools Reused: georss Collection
Instant Browse, Select Datasets/Variables ECHO Mirador NSIDC etc.
Granule OpenSearch, Link to OPeNDAP URLs Space/time Locations
Augmented Cast Linking to Variables Third-Parties Iterate Link to
KML version of cast Re-publish augmented casts Visualize variable,
generate KML Layers A Near Real-Time Mashup Augmenting Event Casts
On-the-Fly LOD
Slide 26
Community Proposal for WG 26
Slide 27
Community Proposal for WG (2) 27
Slide 28
Community Proposal for WG (3) 28
Slide 29
Recommendations 29 Openly Publish All NASA Metadata Collection
& Data Casting, Service Casting, Event Casting Linked Open Data
Continue to Develop Killer Mashups Tools to datasets Datasets to
all relevant services (OpenSearch, WMS, DAP) Science publications
to datasets/variables used Granules to Quality Control annotations
(by DAACs & users) AGU/ESIP People/papers/projects
collaboration graph recommend new collaborators and papers to read
[YOUR App HERE.]