Adding Value While Having Fun With EPA Data

1
Adding Value While Having Fun With EPA Data!

• Brand Niemann
• Co-Chair, Semantic Interoperability Community of
  Practice (SICoP)
• Best Practices Committee (BPC), CIO Council, and
• Enterprise Architecture Team, Office of
  Environmental Information
• U.S. Environmental Protection Agency
• February 23, 2005, and March 2, 2005

2
Data Interoperability Paradigm Shifts

• Organizational
• The National Infrastructure for Community
  Statistics (NICS) Community of Practice (CoP)
  wants to make its data NICS Ready by publishing
  it to the Web in such a way that others can
  easily reuse it! (Like buying a new TV that is
  HDTV Ready.)
• Technical
• The conceptualization of new technical systems
  suffers from technological presbyopia the
  condition of being able to envisage things more
  clearly the farther they are from the present
  realization even though the prospective users
  may grow weary and skeptical while waiting for
  the future to arrive.
• Semantic
• Ontology and ontology patterns are the applied
  use of two basic tenets of software design and
  architecture, indirection and abstraction (see
  Appendix).

3
Overview

• Data is
• Data can be re-purposed
• Data can be mined
• Data can be modeled
• Data can be integrated published
• Data standards can evolve
• Data architecture can be implemented in
  ontology-driven information systems
• Appendix on Indirection Abstraction

This is a semantic approach!
4
Data is

• Unstructured
• Example Web pages
• Semi-structured
• Example Documents (with Table of Contents)
• Structured
• Example Databases (tables)
• Demonstrations in the next section.
• Notes
• This is part of the new Data Reference Model
  (DRM) work.
• But this was there in XML from the beginning -
  everything is a document (all three types).

5
Data can be re-purposed

• Unstructured to Semi-structured
• e.g. Region 4 2004 Press Releases.
• Semi-structured to Structured
• e.g. Region 4 Strategic Plan 2004.
• Structured to Even More Structured
• e.g. TRI data table.
• Demonstrations of Each from the Region 4 Pilot.
• Notes
• We Can Add Metadata and Interoperability At the
  Same Time!
• We Can Even Add Semantic Metadata and
  Interoperability!
• This changes the way we look at portals, document
  and content management systems, registries and
  repositories there really is no difference
  re-purposed data serves all these functions!

6
Data can be mined

• The TRI database is about 8 GB and requires
  industrial-strength tools and analyses for data
  mining, indexing, conversion to XML, and storage
  and retrieval with XML Web Services. This pilot
  demonstrated that large EPA databases can be
  data mined and repurposed into XML
  repositories.
• February 25-26, 2003, Data Mining Technology for
  Military and Government Applications Conference,
  XML Web Services for Data Mining and Repository
  US EPA Toxics Release Inventory, Brand Niemann,
  US EPA, and Data Mining Technology, Jim Walters,
  Insightful Corporation.
• April 9, 2004, Presentations and Demonstrations
  to the EPA Business Intelligence and Analytics
  (BIA) User Group Meeting.

7
Data can be mined
http//www.insightful.com/
8
Data can be mined
I-Miner on TRI 2000 Public Release Data
Pipeline Architecture and Visual Workflow
9
Data can be mined
I-Miner on TRI 2000 Public Release Data
Histogram Plots
Rationale Toxic chemical releases to different
media should be correlated outliers suggest
need to follow up with reporting facilities.
10
Data can be modeled

• EPA Region 4 Pilot
• Phase 1
• October 1, 2004, Implementing a Service-Oriented
  Architecture Pilot Project Design and Initial
  Results.
• October 19 and 25, 2004, Building Enterprise
  Architecture through Web Services for Remote
  Portlets (WSRP) in a Sample EPA Regional Portal,
  Rex Brooks, OASIS, and Ali Naizi, Oracle.
• November 18, 2004, Connecting the Dots CAP-WSRP
  Presented at the XML 2004 Conference as Part of
  the OASIS Interoperability Demos That Showed the
  Integration of Web Services for Remote Portlets
  with the Common Access Protocol and Including
  GeoResponse for Use in the Federal Region 4
  Semantic Interoperability Pilot.
• Phase 2
• February 10, 2005, Semantic Interoperability
  Community of Practice Enablement (SCOPE) for the
  EPA Region 4 Pilot.

11
Current Ontology for Region 4 Pilot
12
Data can be modeled

• The Roadmap
• Re-purpose the 2004 EPA Region 4 Press Releases
  (unstructured to semi-structured) and extract the
  semantic metadata on word usage and frequency and
  use at the top of document.
• Creates a Semantic Web Services node that can be
  integrated with other EPA nodes and Non-EPA
  nodes!
• Compare to what we have now with the EPA Web site
  and the Exchange Network.
• Need ontologies to supplement their topologies
  (topics and Web Services nodes, respectively).

13
Data can be modeled
Demo
Ontology-based search by month, geography, topic,
water semantics.
14
Data can be modeled
Note that search can be focused by both node
hierarchy and/or semantics.
15
Data can be modeled
Ontology (left) is designed by access to the
semantics (right) which is updated automatically
with the addition of new content (not like ETL).
16
Data can be modeled
EPA Web Site
The results is 42 documents out of 533 which then
have to be read and digested!
17
Data can be integrated published

• LandView 6 is integration of
• Census 2000
• EPA EnviroFacts
• USGS Geographic Names Information (GNIS)
  features and
• FGDC Standard Metadata.
• LandView is publishing with
• Interagency collaboration at very low cost
• Public domain software that is easy to use
• Database and software component reuse by design
  and
• Geo-processing (e.g. population estimation) and
  Web-services out of the box (FileMaker Pro)
  (see next slide).

18
Data can be integrated published
Population Estimator
LandView Home Screen
http//www.census.gov/geo/landview/
19
Data can be integrated published
http//www.census.gov/geo/landview/lv6help/pop_est
imate.html
20
Data standards can evolve

• ISO 11179
• EPA Date
• The Date Data Standard provides for a standard
  representation of calendar date in data files for
  data interchange.
• Suggested Upper Merged Ontology (SUMO)
• Date
• According to WordNet, the noun "date" has 8 sense
  (s) (see next slide).
• SUMO is written in the SUO-KIF language
  (declarative semantics and machine processible)
  which has been translated to OWL Web Ontology
  Language.
• See http//www.ontologyportal.org/

21
Data standards can evolve

• Date
• 1. The specified day of the month "what is the
  date today?".
• 2. A particular day specified as the time
  something will happen "the date of the election
  is set by law".
• 3. A meeting arranged in advance "she asked how
  to avoid kissing at the end of a date".
• 4. A particular but unspecified point in time
  "they hoped to get together at an early date".
• The present "they are up to date" "we haven't
  heard from them to date".
• 5. The present "they are up to date" "we
  haven't heard from them to date".
• 6. A participant in a date "his date never
  stopped talking".
• 7. The particular year (usually according to the
  Gregorian calendar) that an event occurred "he
  tried to memorizes all the dates for his history
  class".
• 8. Sweet edible fruit of the date palm with a
  single long woody seed.

22
A Bit of Semantic Humor

• Enterprise Architecture
• Enterprise A Star Trek Spaceship
• Architecture Blueprints
• So, Blueprints of the Spaceship Enterprise!

23
Data architecture can implemented in
ontology-driven information systems

• Ontology-Driven Information Systems
• Methodology Side the adoption of a highly
  interdisciplinary approach
• Analyze the structure at a high level of
  generality.
• Formulate a clear and rigorous vocabulary.
• Architectural Side the central role in the main
  components of an information system
• Information resources.
• User interfaces.
• Application programs.

See for example Nicola Guarino, Formal Ontology
and Information Systems, Proceedings of FOIS 98,
Trento, Italy, 6-8 June 1998.
24
Data architecture can implemented in
ontology-driven information systems

• The Roadmap
• Basic Requirements For an Ontology.
• Generic Process.
• Process for the Indicator Pilot.
• Explanation of the Ontology.
• Schematic of the Ontology.
• Demonstration.
• Semantic Technology Profiles for the Data
  Reference Model (DRM).

25
Data architecture can implemented in
ontology-driven information systems

• Basic Requirements For an Ontology
• 1. Finite controlled (extensible) vocabulary.
• 2. Unambiguous interpretation of classes and term
  relationships.
• 3. Strict hierarchical subclass relationships
  between classes.
• 4. Few others

Source Deborah McGuiness, Ontologies Come of
Age, in the Semantic Web Why, What, and How, MIT
Press, 2002, page 6.
26
Data architecture can implemented in
ontology-driven information systems

• Generic Process
• 1. Identify Scope
• 2. Review Existing Ontologies
• 3. Knowledge Acquisition
• 4. Conceptualization
• 5. Encode
• 6. Test
• 7. Iterate on Steps 3-6.

Source Kathy Lesh, Standards Vocabularies in
Health Care, Kevric, Collaborative Expedition
Workshop 37, NSF, December 9, 2004.
27
Data architecture can be implemented in
ontology-driven information systems

• Process for the Pilot
• 1. Identify Scope Indicators.
• 2. Review Existing Ontologies SUMO-WordNet.
• 3. Knowledge Acquisition Recent GAO Report
  Considered to Be An Excellent Starting Point.
• 4. Conceptualization See Next Two Slides.
• 5. Encode Added to Semantic View of Structured
  Document (See Slide 30).
• 6. Test Inviting Feedback.

Source DRAFT National Infrastructure for
Community Statistics (NICS) Initial Ontology for
Structuring the Community of Practice and Its
Business Case and Products, By Brand Niemann,
January 20, 2005.
Informing Our Nation Improving How to
Understand and Assess the USAs Position and
Progress, GAO, Report to the Chairman,
Subcommittee on Science, Technology, and Space,
Committee on Commerce, Science, and
Transportation, U.S. Senate, November 2004,
GAO-05-1.
28
Data architecture can implemented in
ontology-driven information systems

• Explanation of the Ontology
• Key Terminology
• Source Summary-Background Footnotes
• Topics
• Source Figure 1 and Appendix I
• Level of Jurisdiction
• Source Table 4 of Comprehensive Systems
• Organizational Types
• Source Table 10 of All Systems Studied
  Footnotes

Note Meets McGuiness Requirements in Slide 25.
29
Data architecture can implemented in
ontology-driven information systems
Schematic of the Ontology
Indicators
Topics
Organizations
Jurisdictions
Publicly led
U.S. local/regional level
The Economy
Privately led
U.S. state level
Society Culture
Led by public-private partnership
National level outside the United States
The Environment
Supranational level
Cross-Cutting
Note that each of these classes can and do have
multiple instances underneath them, etc.
30
Data architecture can implemented in
ontology-driven information systems
Demo
See Pilot Projects at http//web-services.gov
Note the folder names are either the ontology or
the knowledgebase instances.
31
Data architecture can implemented in
ontology-driven information systems

• Semantic Technology Profiles for the Data
  Reference Model (DRM)
• Region 4 (See Slide 33).
• Enterprise Architecture FEA-Reference Model
  Ontology (FEA-RMO).
• Taxonomies Formal Ontologies (Michael Daconta
  Recently Published Paper See Next Slide).
• Indicators See previous.
• Community Statistics In process with NICS.
• FHA/NHIN In process with Ontolog Forum.
• ISO 11179 In process with DHS, xmdr.org,
  Ontolog Forum, etc.
• More to be announced as part of the SICoP Module
  3 White Paper Development Implementing the
  Semantic Web.

32
Formal Taxonomies for the U.S. Government

• OWL Listing
• lt?xml version"1.0"?gt ltrdfRDF xmlnsrdf"http//w
  ww.w3.org/1999/02/22-rdf-syntax-ns"
  xmlnsxsd"http//www.w3.org/2001/XMLSchema"
  xmlnsrdfs"http//www.w3.org/2000/01/rdf-schema"
  xmlnsowl"http//www.w3.org/2002/07/owl"
  xmlnsdaml"http//www.daml.org/2001/03/damloil"
  xmlns"http//www.owl-ontologies.com/unnamed.owl
  " xmlnsdc"http//purl.org/dc/elements/1.1/"
  xmlbase"http//www.owl-ontologies.com/unnamed.ow
  l"gt ltowlOntology rdfabout""/gt ltowlClass
  rdfID"Transportation"/gt ltowlClass
  rdfID"AirVehicle"gt ltrdfssubClassOf
  rdfresource"Transportation"/gt lt/owlClassgt
  ltowlClass rdfabout"GroundVehicle"gt
  ltrdfssubClassOf rdfresource"Transportation"/gt
  lt/owlClassgt ltowlClass rdfabout"Automobile"gt
  ltrdfssubClassOfgt ltowlClass rdfID"GroundVehicle
  "/gt lt/rdfssubClassOfgt Etc.

Transportation Class Hierarchy
Source Formal Taxonomies for the U.S.
Government, Michael Daconta, Metadata Program
Manager, US Department of Homeland Security,
XML.Com, http//www.xml.com/pub/a/2005/01/26/formt
ax.html
33
Data architecture can implemented in
ontology-driven information systems

• Additional Demos
• Building Enterprise Architecture through Web
  Services for Remote Portlets (WSRP) in a Sample
  EPA Regional Portal (Oracle 10g)
• Use of Common Alerting Protocol (CAP)
• For example Real-time water quality monitoring
  data from Florida.
• GeoResponse.com
• Voice GIS Multimodal Notification,
  Distributed XML Web Services
• For example January 6, 2005, Norfolk Southern
  Graniteville Train Derailment Scenario.
• Emergency Response Architecture
• Semantic Web Applications for National Security
  (SWANS) Conference, April 7-8, 2005, Trade Show
  Demonstration.

34
Appendix on Indirection Abstraction

• Ontology and ontology patterns are the applied
  use of long-time, fundamental engineering
  patterns of indirection and abstraction.
• Chapter 7 in Adaptive Information Improving
  Business Through Semantic Interoperability, Grid
  Computing, and Enterprise Integration, Pollock
  and Hodgson, Wiley Inter-science, 2004.

35
Appendix on Indirection Abstraction

• Selected tidbits
• Ontology is simply the enabler for software
  engineers and architects to apply core problem
  solving patterns in new and innovative ways.
• Indirection is a concept that is used to plan for
  future uncertainty.
• Simply put, indirection is when two things need
  to be coupled, but instead of coupling them
  directly, a third thing is used to mediate
  direct, brittle connections between them.
• By leveraging indirection in the fundamental
  aspects of the technology, semantic
  interoperability is built for change, and this
  built-in flexibility differentiates semantic
  technologies from other information-driven
  approaches.

36
Appendix on Indirection Abstraction

• Architects of both software and physical
  structures routinely use the principle of
  abstraction to isolate complex components and
  reduce the scope of a problem to be solved (see
  the forest for the trees). By definition,
  ontology is abstraction and is the ultimate
  abstraction tool for information.
• Example Imagine a scenario of using a pivot data
  model without abstraction it would require the
  aggregation of all of the data elements in a
  particular community the result could be the a
  community of 500 applications, each application
  with approximately 100 data elements, requiring a
  pivot model with about 50,000 data elements an
  abstracted model could conceivably be capable of
  representing this information in far fewer than
  about 100 data elements!
• See Demonstrations of SICoP Pilot Projects for
  EPA Managers, August 16, 2004, Semantic
  Information Management (Unicorn) Integrating
  Health and Environmental Information to Protect
  American Children, at http//web-services.gov