Semantic Representation
in Citizen Science
Markus Steinberg
Semantic Web Technologies Seminar
WS 17/18
Citizen Science?
"...the public involvement in inquiry and discovery of new scientific knowledge" - scistarter.com
Citizen Science
A practical example
Protein Folding
Source: Wikipedia - Protein folding
Citizen Science
A practical Example:
Protein Folding
- Costly to automate
- humans "can often see the solution intuitively"
Gamify!
In case you're interested:
scistarter
Citizen Science
Increasing importance
- Internet bringing people and projects together
- Large number of smartphones
- Increasing number of sensors on smartphones
- Lots of independent projects
- Different repositories to share projects and data with the community
Citizen Science
Problem
Goal: Combination & Interoperability
Need for standardization
Citizen Science
Problem
Major Citizen Science organizations have recognized the need for standardization
- CSA
- ECSA
- ACSA
- COST
This presentation...
- ...is about current approaches
- ...presents relevant standards from the semantic web
Geospatial Information
"I saw a hummingbird in California"
"Okay.."
"I saw a hummingbird in South Africa!"
"Oh wow, awesome!"
Geospatial Information
Open Geospatial Consortium
- Developing interface standards for geospatial information
- Lots of standards
-
Especially important: GeoAPI 3.0
- Java interface
- Types and methods for manipulation of geographic information
Provenance Information
- Lots of unknown people involved in data collection
- What about the quality? Reliability? Reproducibility?
Provenance Information
PROV-Ontology
Who did what in which way, on whose behalf,...?
W3C
Provenance Information
PROV-Ontology
Source: PROV-O Documentation
Sensor Descriptions
- Different smartphone sensors
- Specialized equipment
SSN & SOSA
-
Semantic Sensor Network (SSN)
- Includes Sensors, Observations, Sample and Actuator (SOSA) Ontology
W3C
Sensor Descriptions
SSN & SOSA
Sensor Descriptions
Source: SSN & SOSA Documentation
Persons
Lots of persons with relationships to each other, to projects, to organizations, ...
Friend of a Friend (FOAF)
Metadata...
- ...of projects
- ...of datasets
Source: http://211.185.62.34/
Metadata
Dublin Core
- Metadata Element Set- 15 core elements
- DCMI Metadata Terms (dcterms)
Metadata
Metadata Element Set
Allows to describe:
- Title of a resource
- Creators, contributors, publishers
- Descriptions
- Topics
- ...
Everything you need for a basic search
Metadata
dcterms
- Properties with formal domains and ranges
- "Refined terms"
- References to external vocabularies
- References to syntactic standards
Built on top of the Element Set
Metadata
dcterms
- Intended audience
- Date of creation, publication, modification
- References to other versions
- ...
Observations
- What was observed
- Value of the measurement
Observations
OBOE
Extensible Observation Ontology
- Originally developed for biodiversity research
- Seems to be under active development
Observations
OBOE
Observations
OBOE
Extensible Observation Ontology
OBOE encourages extension:
- Defining subclasses of Entity
- Using units & characteristics from other ontologies
Alignments
- Some ontologies and standards include similar terms
-
Alignments describe mappings between these similar terms
- Ensures compatibility
Alignments
- SSN/SOSA - OBOE
- SSN/SOSA - PROV-O
- SSN/SOSA - O&M
- DCMI - PROV-O
SWE4CS
OGC - Citizen Observatory Web (COBWEB)
- Joint effort of different CS observatories
- Goal: Standardized way of collecting and modeling CS observations
- Based on existing standards
SWE4CS
- Core: Observations & Measurements Standard
- SensorML
- SWECommon
- ISO 19109
Based on...
Design decision: Completely base the model on these existing standards without adapting to CS needs
SWE4CS
- Name & description of an observation
- Observed property
- Location information
- Time
- Quality
- Procedure & protocol
- Persons & parties
- Collections & aggregations of observations
What can be described?
PPSR_CORE
- Program Data Model Metadata Standard / data sharing protocol
- In development by CSA
- Joint effort of different CS repositories
- Declared objective of reusing existing standards
Public Participation in Scientific Research
PPSR_CORE
Common Data Model
PPSR_CORE
Project Data Model [WIP]
- Project name, aim and description
- Start date
- Geographical information
- Activity status
- Responsible organization
Required fields (Selection)
PPSR_CORE
Project Data Model [WIP]
- Tags, keyword, topics
- URL
- Contact & Funding information
- ...
Optional fields (Selection)
PPSR_CORE
Project Data Model [WIP]
- PROV-O
- OGC GeoAPI
- FOAF
Mappings
PPSR_CORE
Dataset Data Model [WIP]
- Identifier, name, abstract
- Access rights, licenses
- Geographical information
- Status
Required fields (Selection)
PPSR_CORE
Dataset Data Model [WIP]
- Publication and download
- Update frequency
- ...
Optional fields (Selection)
PPSR_CORE
Dataset Data Model [WIP]
- Dublin Core
- GeoAPI
Mappings
PPSR_CORE
Current development
- No structured publication (RDFS, ...)
- Focus on Common Data Model
- Analyzing Use Cases
- Discussions with stakeholders
- Evaluating existing standards
Conclusion
- PPSR_Core seems most promising
-
Can a one-size-fits-all approach even work?
- Very different requirements for different projects and organizations