Implementing the Open Government Directive using the technologies of the Social Data Web

Post on 11-May-2015

2,260 views 1 download

Tags:

description

This presentation demonstrates the use of Semantic Web technologies with Social Networking tools, considering metadata specifications as Social Media. Example ontologies and instance data from the Capital Planning and Investment Control and Business Motivation are created that link 'what' (Agency IT investments) with 'why' (Agency goals and objectives), using a simple linking ontology. Knowledge Workers use a Semantic Halo Mediawiki to curate the data.

transcript

Implementing the Open Government Directive

using the technologies of theSocial Data Web

George Thomas, December 2009

2

This Presentation

• Demonstrates the value and utility of the Social Data Web

– Using two different Websites (part 1)

• One to host a few simple ontologies (metadata specifications)

– Using Semantic Web and XML standards based technologies

• And another to host their corresponding (instance) datasets

– With a Semantic MediaWiki (part 2)

• That imports both the ontologies and datasets off those 2 Web sites

– For easy human interaction - creation, curation and augmentation

• But maintains the independence of each domain ontology/dataset

– Making machine readable representations available on demand

• And proposes a HHS Open Government Directive project

– Realizing next-gen Performance Management

• By linking Mission Goals (why) with IT Investments (what) data

3

W3C eGov IG - Social Data Web Project

4

Metadata Specs Published on the Web

• 3 domain specific ontologies– /vcso/omg/bmm

• The OMG Business Motivation Metamodel (BMM)

– /fedsheets/omb/ex300 and ex53 • OMB Federal Capital Planning and Investment Control (CPIC)

• 1 integration or ‘bridge’ ontology– /bizmo (that connects CPIC and BMM)

5

• These datasets are instances of– http://ontologies.socialdataweb.org domain specific and

integration ontologies

• Simulating 3 different agency.gov/open data sources– Each builds on, and adds to the others, without changing them

Instance Data Published on the Web

6

it.usaspending.gov -> HHS Ex300

• Selecting columns to populate CSV

7

Spreadsheet->Ontology Pattern

• Get (raw?) data

– From http://it.usaspending.gov CSV download

• Worksheet name = Class name

– For example, an ex53 or ex300 spreadsheet

• ColumnName = Class Datatype Property

• Row = instance of Class

– First column (often) treated as record (row) PK

• Cells = instances of Datatype Properties

8

/fedsheets/omb/ex300/ex300.owl#

• Version 1.1 (shown here)

– 1 Class

– 25 Datatype Properties

• With XSD typed literal values

– Replaced xsd:string where obvious

• With xsd:float, xsd:int, and xsd:date

• Version 1.2 (future)

– Should refactor further

• Introduce more (obvious) types

• Reuse authoritative/defacto vocabs (!)

– And create more object props

• That link types to other types

9

OMB Ex300 Metadata Specs on the Web

10

HHS Ex300 Instance Data on the Web

11

/fedsheets/omb/ex53/ex53.owl#

• Version 1.1 (shown here)

– 1 Class

– 17 Datatype Properties

• With XSD typed literal values

– Replaced xsd:string where obvious

• With xsd:float

• Version 1.2 (future)

– Should refactor further

• Introduce more (obvious) types

• Reuse authoritative/defacto vocabs (!)

– And create more object props

• That link types to other types

12

OMG BMM Simplification

• Object Management Group (OMG) Business Motivation Metamodel (BMM)

– A desired result is a generalization of goals and objectives • A goal is something an Org is trying to

achieve• An objective quantifies a goal, specifying

timing and measurement– A course of action is something an Org

does to achieve a desired result• A strategy is a broad, lasting course of

action• A tactic is a narrow, fleeting course of

action– An influencer is something that can

affect the Org’s ability to achieve its goals or implement its strategies• An actuator is an influencer that can be

considered as a quantity that can increase or decrease over time

– An assessment is a judgment of an influencer’s affect on an Org

13

Full BMM Ontology in TBC-FE

• 51 Object Properties

<owl:ObjectProperty rdf:ID="focusesEffortOn">

<rdfs:domain rdf:resource="#Strategy"/>

<rdfs:range rdf:resource="#Goal"/>

</owl:ObjectProperty>

• 35 Classes

<owl:Class rdf:ID="Strategy">

<rdfs:subClassOf rdf:resource="#CourseOfAction"/>

</owl:Class>

<owl:Class rdf:ID="Goal">

<rdfs:subClassOf rdf:resource="#DesiredResult"/>

</owl:Class>

– RDF/XML above, N3 below:Goal

a owl:Class ;

rdfs:subClassOf :DesiredResult .

14

OMG BMM MetaModel (=Ontology)

15

OMG BMM Metadata Specs on the Web

16

BMM Legend: Means to End, Assets

• A ‘business view’ of the Business Motivation Metamodel (BMM)

• Let’s look at some BMM instance models and resulting data now!

17

CCAC BMM Ends: Instance Model

• http://govgoals.org/data.cio.gov/open/cloud/ccac/ccac.owl

18

CCAC BMM Ends: Instance Data

19

AIC BMM Ends, Means and Assets: Instance Model

20

AIC BMM Ends, Means and Assets: Instance Data

21

HHS + AIC + CCAC BMM Instance Model

22

HHS + AIC + CCAC BMM Instance Data

23

Bizmo ‘Bridge’ Ontology in TBC-FE

24

KR Benefit: Subsumption Inferences

25

Key /bizmo.owl#Classes

26

Key Interlinking /bizmo.owl#properties

– #supports properties link CPIC ex53/ex300’s• to BMM Strategies, Tactics, Goals, Objectives, etc.

– extending the Ex53/300 specs – without changing them!

– #maintainsExhibit53/300 link OrganizationalUnit’s• to CPIC investment information

27

Bizmo Metadata Specs on the Web

28

End of Part 1

• Recap:– We created and published 2 CPIC ontologies

• We created and published HHS instance datasets of each– Using data sourced from it.usaspending.gov, the ‘ITDB’

– We created and published 1 BMM ontology• We created and published CCAC, AIC and HHS instance datasets of

the BMM– Mostly on the subject of Cloud Computing

– We created and published a ‘Bizmo’ bridge ontology• We created and published Bizmo instance datasets that link HHS CPIC

(what) datasets to a variety of CCAC and AIC BMM (why) datasets

• What now?– Let’s interact with all this!

• Using the ubiquitous, free, open source and arguably the most successful social networking tool ever!– MediaWiki, the software that powers Wikipedia.org

29

Paul Allen’s Project Halo

30

Ontoprise SMW + Halo

31

Browsing SMW KB via Classes

32

Browsing SMW KB via Properties

33

Ontology Classes/Properties = Tags

• Create semantic annotations (part of wysiwyg editor tools)– auto-completion suggests tags to reuse from ontologies that have

been imported into the wiki (note existing tags from BMM ontology)

34

Ontology based SMW KB Query

35

SMW KB Report Generation

36

Exhibit300 ‘Facts’ about HHS UFMS

37

NHIN – Automagic Data Merging

38

NHIN RDF/XML IEP: Export Excerpt

• Wiki instance data and metadata curation (SME edits) maintain class and property specs of ontologies imported off the Web when exported or accessed by other sites/users/apps

39

Browsing Linked Datasets on SMW

• Filtering through properties defined by Bizmo ontology• finds CPIC instance data that is linked to BMM instance data

40

SDW = SNS enabled LOD• This presentation is summarized by the interactive data below• And - it’s a Wiki – SME’s can easily add annotations and data!

End of Part 2

• Recap:– We imported the ontologies and datasets on the Data

Web (from Part 1) into a Semantic MediaWiki (SMW)– We interacted using ‘Halo’ SMW tools by;

• Browsing the instance data via ontology classes and properties (aka ‘faceted browsing’)

• Curation of imported instance data and adding semantic annotations, creating new metadata and instance data

• Querying the wiki using ontology classes and properties and instance data Web ID’s (aka URI’s)

• Navigating from ‘facts’ about any datum to any other• Extracting data from the Wiki in a variety of machine and

human readable output formats

Conclusion

• This presentation has demonstrated an implementation of;– Social Data Web features and functionality– That links Business Strategy with IT Investment data– Using existing free and open source software tools– To design ontologies (metadata schemas) that enable– Federated instance datasets to be integrated– Treating both metadata & datasets as social media– All of which are implemented on the Data Web– Without writing one line of application code

Proposal

• Continue this SDW work– Under the auspices of the HHS Innovation Council

• And continuing with the W3C eGov IG

– Targeting Federal CIO Council and OMB Ends• Stand up SDW as stated in AIC Cloud Computing proposal• Demonstrate next-gen IT Dashboard functionality

– As part of HHS OGD response plan

• Next steps– Add OMG FEA and MPG metamodels and defacto vocabularies

(DC, FOAF, SIOC, etc.) and DSL’s (risk, security, etc.)• Refactor example ontologies accordingly

– Stand up GovGOALS using SMW+Halo on public cloud (move off my laptop )

• Create/demonstrate syndication/dissemination capabilities

Thanks For Your Time and Consideration!

• Contact me:

– http://xri.net/=george.thomas

• About me:

– HHS Enterprise Architect• CTO Council, member• Innovation Council, member

– Federal CIO Council • Architecture and Infrastructure Committee (AIC), Services Subcommittee co-chair• Cloud Computing Advisory Committee (CCAC), member

– W3C eGov IG, Invited Expert

– OMG GovDTF, Steering Committee member

– Graduate School, SOA Faculty Instructor