ourSpaces: a Collaborative Research Environment for eSocial Science - PowerPoint PPT Presentation

Title: ourSpaces: a Collaborative Research Environment for eSocial Science


1
ourSpaces a Collaborative Research Environment
fore-Social Science

• Feikje Hielkema, Peter Edwards, Alison Chorley,
• Edoardo Pignotti, Chris Mellish, Alun Preece
• John Farrington
• Computing ScienceGeography Environment
• University of Aberdeen

Funded under the UK ESRC eSocial Science
programme award reference RES-149-25-1027.
2
Content

• Introduction
• e-Social science
• Virtual research environments
• Web2.0
• ourSpaces
• Requirements
• Prototype
• Feedback
• VRE components
• Metadata infrastructure
• Natural language interface
• ourSpaces services
• Discussion

3
Background

• eSocial Science
• The use of advanced computing technologies to
  facilitate sharing of distributed and
  heterogeneous data and compute resources, plus
  enhanced collaboration support, for social
  scientists.
• Social Science challenges for eScience
• Heterogeneous data types documents, video,
  audio, survey instruments, annotations, numerical
  data, secondary and/or primary data sources.
• Diverse approaches to data collection and
  analysis.
• Working across traditional disciplines.
• Methodologies not necessarily linear.
• Practical and epistemological considerations when
  sharing and re-using data.
• Increased emphasis on multi-method or
  mixed-method approaches, where emphasis is placed
  on plural types and sources of data.

4
PolicyGrid eSocial Science

• National Centre foreSocial Science
• Investigates specific applications of eSocial
  Science, develops tools to support them and
  advises on the future strategic direction of
  eSocial Science and eScience.
• Coordinating eInfrastructure effort for UK ESRC.
• www.ncess.ac.uk
• PolicyGrid
• One of seven ESRC-funded research nodes, all of
  which involve Computer Scientists and Social
  Scientists.
• www.policygrid.org

5
PolicyGrid Aims

• To explore how a range of evidence types can be
  integrated to support evidence-based policy
  development and appraisal
• To investigate flexible metadata frameworks and
  their role in enhancing the ability of social
  scientists to communicate/share results
• To develop novel interfaces and tools which
  enable social scientists to utilise Grid-enabled
  resources and to embed them within their existing
  working practices
• To enhance workflow tools to improve
  documentation and control of workflow
  experiments, thus allowing such experiments to
  form part of policy evidence bases
• To promote awareness of the Semantic Web/Grid
  vision and supporting technologies amongst social
  scientists.

6
Virtual Research Environments

• What should a VRE do?
• Should support the processes of research (could
  include administration)
• Should be built on interoperable tools (hence be
  extensible, capable of being tailored, etc.)
• Should support a wide range of users and user
  requirements, and be compatible with existing
  widely-used tools.
• Specific examples
• Access to computational resources, data storage,
  remote instrumentation, electronic information.
• Literature search and retrieval, support for
  authoring, e-prints/self-archiving.
• Messaging (including secure messaging), sharing
  diaries/calendars/files etc., distributed
  document production.

UK JISC VRE programme meeting, January 2005.
7
Virtual Research Environments

• FearlusG MyWorkspace
• Provides access to a Grid-enabled social
  simulationmodel of land-use change.
• Metadata tools support annotation/sharing
  ofsocial simulation resources experiments,
  parametersets, etc.
• OWL ontology of scientific resources.
• Simple support for construction of scientific
  arguments.
• http//www.csd.abdn.ac.uk/research/fearg/

8
Virtual Research Environments

• MyExperiment a workflow bazaar
• Contribute to pool of scientific workflows
  share,re-use and repurpose them.
• Share expertise.
• Build communities form relationships.
• http//myexperiment.org/
• Integrative Biology IBVRE
• A graphical, online environment to
  allowbiological simulation experiments to
  beconstructed and managed.
• Uses CCLRC scientific metadata model.
• Extensive user requirements and design activity.

9
Virtual Research Environments

• VERA
• Virtual Research Environment for Archaeology
• Aims to produce a fully-fledged virtual
  researchenvironment for the archaeological
  community.
• Documenting archaeological excavation
  andassociated finds.
• Builds on earlier Silchester 'Town Life project.
• http//vera.rdg.ac.uk

• Support for community based activities
  including collaboration - a common theme in
  VRE development.
• Community-driven trend in Web technologies.
• The Social Web or Web 2.0

10
Web 2.0

• Services such as social networking sites, wikis,
  communication tools, and folksonomies that
  emphasize online collaboration and sharing among
  users.
• Themes
• collective intelligence
• users as co-developers not just end users
• architecture of participation
• distribution rather than centralisation.

http//web2.wsj2.com/
11
BibSonomy
12
(Initial) Requirements for an eSocial Science VRE

• Through our work as part of the PolicyGrid
  project, we have identified the following
  desirable features for a VRE
• support for management of datasets
• tools for annotation of resources
• must acknowledge diversity of social scientists,
  their methods and working practices
• metadata support must be lightweight, flexible
  and community-driven
• integration with desktop software tools.

13
ourSpaces - Prototype

• A collaborative working environment designed to
  enable social scientists to work together and to
  share and reuse resources in the context of the
  eSocial Science Grid.
• Integrates aspects of Grid, Semantic Web and Web
  2.0 technologies.
• Support for
• Social networking.
• Resource management - upload, search, annotation,
  etc.
• Creation of project spaces.
• Publishing blogs and wikis.
• Execution/monitoring of social simulation
  workflows.
• Integration with other tools (e.g. Squanto
  qualitative analysis desktop tool).

14
ourSpaces - Prototype

• Environment built upon a flexible metadata
  framework, supporting resource annotation,
  provenance, collaboration.
• Integrates a natural language driven mechanism
  for creating, querying and browsing metadata.
• Uses the concept of a space as a way of
  organising resources and tools.
• Integrates, but does not attempt to replace,
  existing tools.

15
ourSpaces

• The concept of a space is central to the VRE.
• Two types of space
• Resource Space
• A collection of resources organised by type
  documents (inc. sub-categories) simulation
  models maps, etc.
• Space contains relevant software/service tool set
  for those resources.
• Community Space
• An area which represents a virtualorganisation
  (community) sharinga common interest.
• Within a space - a common taskset, terminology
  (expressedthrough folksonomies), Gridtool set.

16
Prototype Feedback

• Existing prototype created using HTML and JSP -
  to act as a platform for gathering further user
  requirements.
• Feedback from social scientists
• Managing access rights VERY important
• Choose your spaces different disciplines are
  interested in different services/have different
  vocabularies
• Project space has universal appeal!
• use ourSpaces to create project website for
  general public
• project members share resources and information.
• Great interest in collaborator list
• separate into different types of collaborators
• social networking applications view your
  collaborators collaborators.

17
Prototype Feedback

• Integrate services into ourSpaces, making it a
  central place for research activities.
• Dont make this yet another thing I have to log
  into.
• Some of the wilder ideas
• peer-review of resources
• create your CV with ourSpaces
• structure your research with an ourSpaces desktop
  tool.

18
Metadata Infrastructure

• ourSpaces utilises several OWL ontologies to
  capture
• metadata describing a resource.
• metadata describing the process used to create a
  resource.
• high level annotations about a process.
• metadata about the process of constructing
  evidence assertions.
• Ontologies have been aligned with the schema used
  by the UK social science data archive (UKDA),
  itself based on the international DDI standard.

19
Metadata Infrastructure

• ourSpaces is trying to use a community
  environment to create and add value to social
  science resources, and to promote their re-use
  (when appropriate).
• Tom Gruber (ISWC 2006)
• http//iswc2006.semanticweb.org/program/keynote_gr
  uber.php
• Suggests two approaches to integrate ontologies
  and Web 2.0 style tags.
• Construct an ontology of tag data - to support
  integration of user-contributed data across
  applications.
• Merge structured data (ontology) with
  unstructured content (tags) - create aggregate
  value and provide context for tags.

20
Metadata Infrastructure

• ourSpaces integrates utility ontologies with
  folksonomies.
• Ontologies only provide context (and an
  organisational structure) within which tags
  provide descriptive information.
• Find me all datasets collected by John Farrington
  whose data were gathered in Wales and which used
  the APAT methodology.
• User community drives a folksonomy for each
  rdfdatatype property these may be specific to a
  certain space.

21
Providing Access to Metadata Usability Issues

• We want to provide a consistent interface for all
  metadata related tasks creation, querying and
  browsing.
• Minimise learning time and effort on part of
  users.
• Users should not have to switch tools when they
  switch tasks.
• How to present metadata?
• Graphical readership is an acquired skill for
  novices, natural language is faster and less
  prone to interpretation errors (Petre, 1995).
• Social science is not a discipline in which this
  graphical readership skill is necessarily
  acquired.
• Use of natural language to present and elicit
  metadata.

22
Natural Language Interface

• How to let user create metadata?
• Parsing all user utterances beyond
  state-of-the-art.
• Existing approaches therefore restrict user
  input
• Natural Language Menus
• Controlled Languages
• WYSIWYM What You See Is What You Meant(Power et
  al., 1998)
• System generates a text presenting all the
  information the user specifies.
• The user edits the text to add more information.
• Driven by underlying data structure (in our case
  an ontology integrated with folksonomies).
• Used for query formulation as well as document
  authoring.
• Positive evaluation.

23
Natural Language Interface
24
Natural Language Interface
25
Natural Language Interface
26
Natural Language Interface
27
ourSpaces Services

• A range of underlying Web and Grid services
  provide functionality for users and support VRE
  administration.
• Data Management
• Digital object store (based on Fedora) provides
  repository for uploaded resources (documents,
  data files, audio, etc.)
• RDBMS enabled as a Grid service (via OGSA-DAI).
• Metadata
• High performance RDF repository service (Sesame).
• Tag service (utilising RDBMS) manages tag
  data,folksonomies - by RDF property and context
  (space).
• Simulation
• FEARLUS land-use model service.
• Services based upon Netlogo and Repast tools.

28
Architecture
Grid Middleware
Fearlus
Repast
Netlogo
Workflow
Simulation Services
Simulation Models
Reasoning Services
Tag service
Sesame
Data Access Services
RDBMS
29
Discussion

• ourSpaces Web 2.0
• metadata driven by folksonomies ontologies.
• community-driven content, resource creation and
  sharing.
• our spaces - so users must be able to
  configure/re-purpose community spaces as they
  deem appropriate.
• a space as a tool/service mash-up.
• Evaluation of current prototype ongoing.
• www.ourspaces.net survey.
• refinement of requirements.
• small user group to be engaged in participatory
  design.
• Separate evaluation studies of different
  ourSpaces components.

30
Discussion

• ourSpaces should evolve to allow components from
  other eSocial Science efforts to be integrated,
  e.g. GeoVUE tools within Map space, MoSeS model
  within the Models space.
• Ongoing ourSpaces development will be resourced
  through PolicyGrid and the wider NCeSS
  eInfrastructure activity.

31
WANTED

• PolicyGrideInfrastructure Position
• To focus on the development of Grid services and
  tools in the context of the metadata simulation
  workpackages of the wider UK ESRC eInfrastructure
  effort.
• Includes a Grid service based metadata
  infrastructure, portal development, novel portlet
  solution for RDF creation, querying, etc.
• 27,857-33,262 p.a.
• 21 months full-time
• Available immediately.

Contact Pete Edwards - p.edwards_at_abdn.ac.uk
32
Questions?