presenting provenance based on user roles

Post on 24-Feb-2016

30 Views

Category:

Documents

0 Downloads

Preview:

Click to see full reader

DESCRIPTION

Presenting Provenance Based on User Roles. Experiences with a Solar Physics Data Ingest System. Patrick West, James Michaelis, Peter Fox, Stephan Zednik, Deborah McGuinness – Tetherless World Constellation (http://tw.rpi.edu) – Rensselaer Polytechnic Institute (http://www.rpi.edu). - PowerPoint PPT Presentation

TRANSCRIPT

1

Presenting ProvenanceBased on User Roles

Experiences with a Solar Physics Data Ingest System

Patrick West, James Michaelis, Peter Fox,Stephan Zednik, Deborah McGuinness – Tetherless World Constellation (http://tw.rpi.edu) – Rensselaer Polytechnic Institute (http://www.rpi.edu)

AGUFM2010-IN43C-05

2

Outline of Presentation

• Prior Work in Selective Provenance Presentation• Rationale for User Roles in Presentation• Our Focus Area:

• Semantic Provenance Capture in Data Ingest Systems (SPCDIS)

• Advanced Coronal Observing System (ACOS)• Applying user roles to provenance

3

Prior Work• Significant prior work on provenance views +

abstractions (Moreau, 2009)• Two kinds approaches:

• Expanding Abstract Provenance (Hunter, 2007)• Start with abstract provenance, expand to fine

grained• Abstracting Fine Grained Provenance (Davidson,

2008)• Start with fine-grained, select desired

components, then abstract away unwanted detail

• Common goal: manage complexity of provenance

4

Complexity

5

Kinds of Users

• In context of a Solar Physics Data System, two kinds of expertise:

• Scientific (Astro/Solar Physics)• Technical (Pipeline + components)

• Kinds of Users:• Project coordinators

• Knowledgeable in both science and technical• Outside Domain Experts• Citizen Scientists

6

Rationale for User Roles• Different backgrounds for different users

• E.g., Domain Experts versus Citizen Scientists• Abstract -> Fine-grained: can be time intensive

process• Fine-grained -> Abstract: requires background to

know what you’re looking for• Key idea: Initial presentation of provenance

components can be important for end-users• Finer grained components for experts• Abstract components for novices

Multiple-domain knowledgebase

• Objective: Use Semantic Web technologies to combine provenance from different sources in an interoperable fashion.

7

Provenance Ontology

Solar Physics Domain

DataProcessing

Domain

Extension of work on Virtual Solar

Terrestrial Observatory

http://www.vsto.org

Good/Bad/Ugly (GBU) ratings,

Trust, Quality flags

Proof MarkupLanguage (PML)http://inference-web.org

8

Advanced CoronalObserving System

Mauna Loa Solar Observatory (MLSO)Hawaii

Intensity Images (GIF)

• Raw Image Data

Raw Image DataCaptured by CHIPChromosphericHelium-I ImagePhotometer

• Raw Data Capture

National Center for Atmospheric Research (NCAR) Data Center.Boulder, CO

Velocity Images (GIF)

• Follow-up Processing on Raw Data • Quality Checking (Images Graded: GOOD, BAD, UGLY)

Publishes

9

Provenance View –Citizen Scientist

Data Capture (MLSO)

Data Processing (NCAR)

Quality Check (NCAR) Good/Bad/

UglyRating

• Raw Image Data

• Calibrated Image Data

10

Provenance View –Domain Expert

Data Capture (MLSO)

Flat Field Calibration

Good/Bad/Ugly Rating

Hot Pixel Correction

Centering/Trimming/Clipping

Compute Sample Means

Determine Test Channel

Assign GBU Rating

Data Processing

Quality Check

11

Use Cases• Different users wish to get overview of provenance

for quality rating.• Citizen Scientist:

• Sees high-level provenance.• Wishes to know more about how Good/Bad/Ugly rating

created• Expands Quality Check node.

• Domain Expert:• Starts with fine-grained provenance view,

generates abstraction exposing quality check processes:

• Compute Sample Means• Determine Test Channel• Assign Good/Bad/Ugly Rating

12

Applying user roles

• Semantic Web (RDFS/OWL) Ontologies for defining domain knowledge needed. Specifically for defining:• Workflow components.• User roles.• Component-Role Mapping.

RDFs – Resource Description Framework schemaOWL – Web Ontology Language

13

Ongoing Issues• Some inherent challenges

• Deciding on how to map components to roles.• Will a given user necessarily fit into one of the pre-

defined roles?• Key research question pursued

• For preserving provenance interface usability, what a good middle ground between:• Going from abstract to fine-grained provenance• As well as fine-grained to abstract provenance

14

Summary• Managing complexity is an important activity for

presenting provenance.• Just providing drill-down from abstract to more

detailed views or fine-grained selection is not enough.

• The user can be provided an initial presentation of content based on their level of knowledge, from general interest to domain expert.

• What is needed is an approach that provides the right level of initial explanation based on the user’s role.

15

References• L. Moreau, 2009. “The foundations for provenance

on the web.”• K. Cheung, J. Hunter, and Lashtabeg, A. and J.

Drennan “SCOPE: a scientific compound object publishing and editing system.” International Journal of Digital Curation, 3(2), 2008.

• S. Cohen-Boulakia, O. Biton, S. Cohen and S. Davidson “Addressing the provenance challenge using ZOOM.” Concurrency and Computation: Practice and Experience, 20(5), p. 497-506, 2008.

top related