sonet: scientific observations network semtools: semantic enhancements for ecological data...

34
SONet: Scientific Observations Network Semtools: Semantic Enhancements for Ecological Data Management Mark Schildhauer, Matt Jones, Shawn Bowers, Huiping Cao

Upload: jessica-allison

Post on 30-Dec-2015

220 views

Category:

Documents


0 download

TRANSCRIPT

SONet: Scientific Observations Network

Semtools: Semantic Enhancements for Ecological Data Management

Mark Schildhauer, Matt Jones, Shawn Bowers, Huiping Cao

Outline

• Project overview• Observational data models (SONet)

• OBOE re-factored• O&M• EQ• SONet core - domain ontologies

• Semantic tools for observational data (Semtools)• Semantic annotation• OM Query language• Querying annotated observational metadata and data

Challenges

• Mark and Matt fill in …

Our Approach

The SONet Project- Define a core observational model (based on OBOE, O&M,

and others) … progress made on OBOE/O&M alignment- Identify and develop domain-specific vocabularies for

describing observational data semantics- Define a set of scientific use cases for data interoperability- Develop a set of interoperability “demonstration prototypes”

The Semtools Project- Incorporate OBOE and semantic annotations into existing

Metacat and Morpho metadata tools- Focus on ecology data and use cases

Weight Mass Unit

Biomass

usesStandard

Gram

is-a

Bio.EntityTree

Leaf LitterTree Leaf Wet Weight Dry Weight

Observation

Measurement

Site Species Ind Mass

GCE6 Picea Rubens 1 75.13

GCE6 Picea Rubens 2 179.81

GCE7 Picea Rubens 1 443.20

… … … …

Data

Structural Meatadata

<attribute id=“att.4”> <attributeName> Mass </attributeName></attribute>

hasMeasurement

OBOE Semantic Annotation

Domain-Specific Ontology

is-a is-ais-a

is-a

has-part

hasCharacteristicis-a

part-of is-a

ofEntity

usesS

tandardofCharacteristic

Observational data models

• OBOE 1.0• O&M (ISO)• EQ• SONet Domain Ontologies

• Plant Traits• SBC• Ongoing

OBOE Core 1.0

Extensible Observation Ontology (OBOE) [1]

7

Entity

Characteristic

Observation

Measurement

Protocol Standard

hasMeasurement

ofCharacteristic

usesProtocol usesStandard

ofEntity

hasContext

1..1*

1..1

*

*

*

1..1 1..1

1..1**

hasValue

1..1

**

OBOE re-factored (cont.)

• Shawn adds more (2-3 slides) on OBOE …

O&M ([2])

9

• Feature– Abstraction of real world phenomena

(Def. 4.5)– E.g., Tree

• Feature type– Class of features having common

characteristics (Def. 4.6)• Property

– Facet or attribute of an object referenced by name (Def. 4.14)

– E.g., Height• Property-type

– Characteristic of a feature type (Def. 4.15)

Feature

Property

carrierOfCharacteristic

Feature

Property

OM_Observation

observedProperty

featureOfInterestcarrierOfCharacteristic

OO&M (Cont.)

• Observation– Act of observing a property (Def. 4.10)

• Measurement– Set of operations having the object of determining the value of a quantity (Def.

4.9)

• OM_Observation– An instance of feature type

Feature

Property

OM_Observation

Result

hasResult

Procedure

usesProcedure

carrierOfCharacteristicobservedPro

perty

featureOfInterest

O&M (cont.)

• Observation procedure– Method, algorithm or instrument, or system of these which may be used

in making an observation (Def. 4.11)– The base class OM_Process

• Observation result– Estimate of the value of a property determined through a known procedure.

Any type

Feature

ObservationContext

Property

OM_Observation

Result

relatedObservation

hasResult

Procedure

usesProcedure

ObservationContext• Some observations depend on other observations to provide context in

understanding the result. (Sec. 6.2.4)

• Link a OM_Observation to another OM_Observation, with the role name relatedObservation for the target

carrierOfCharacteristicobservedPro

perty

featureOfInterest

O&M (cont.)

Entity Feature

Characteristic Property

Observation Measurement OM_Observation

Protocol Procedure

ResultStandardEntity

hasContext ObservationContext

Entity

CharacteristicMeasurement

Observation

Standard

hasMeasurement

ofEntity

Protocol

Feature

ObservationContex

Property

OM_Observation

Result

carrierOfCharacteristicobservedProper

ty

featureOfInterestrelatedObservation

hasResult

Procedure

usesProcedure

13

hasValue

usesProtocol usesStandard

ofCharacteristic

hasContext

EQ ([4])

• Entity– Describes some object in the real world. E.g., eye

• Quality– Describes an entity's attribute and its attribute

• Character – Composed of Entity and Quality attributes to

• Character state– Quality value. E.g., “red” to represent eye’s color is red). - represent the meaning of which entity's which attribute.

E.g.,- eye's color.- value. E.g., color = red, means eye’s color is red.

Comparison

OBOE re-facored O&M EQ

Entity Feature Entity

Characteristic Property Quality attribute

Observation

OM_ObservationMeasurement Quality value or Character state

Protocol Procedure

Standard + Result Result

hasContext ObservationContext

SONet core - domain ontology: trait• Trait group: Centre d'Écologie Fonctionnelle et Évolutive (CÉFÉ) [5]

SONet core -domain ontology: SBC• Group: Santa Barbara Coastal (SBC) Lont Term Ecological Research (LTER) [6]

• Plant– iPlant group ([7])

• PATO ([8])– Phenotypic Quality Ontology– Automatic tool to convert PATO to be compatible with SONet

core model

SONet core -domain ontology: ongoing

How to use SONet core model?• SONet-core

• Morpho data annotation tool to generate data instances for you automatically

• OM query language • OM query framework for data discovery

• O&M• Write your own xml file to contain the observation

and measurement data• Write Schematron [3] file to validate whether the

data xml file • No tool report!

19

Tools

• Morpho data annotation tool to generate data instances for you automatically

• Observation and Measurement (OM) query language

• Framework for querying annotated observational data

20

Semantic Annotation

OM Query example

• Tree[Height > 5 Meter]- Return datasets that have at least one Tree

observation containing a Height measurement with a value greater than 5 Meters

• Tree[Height > 5 Meter], Soil[Acidity >= 7 pH]- Return datasets that contain at least one Tree

observation (having a measurement where the Height was greater than 5) and at least one Soil observation (having an Acidity measurement of 7 or greater)

22

OM Query Example (cont.)

• Tree[Height > 5 Meter] -> Soil[Acidity >= 7pH]- Incorporates context via the "->” (arrow) symbol, which

can be read as "contextualized by" or "has context"- Returns datasets that contain at least one Tree observation

(with the corresponding height value) where the observation was taken within the context of a Soil observation (with the corresponding acidity value)

23

Query framework

Annotation 1 Annotation 2 Annotation 3 Annotation 4

File 1 File 2 File 3 File 4

Offline data processing

Result Q1 Result Q2 … ….

OBOE-aware DB

…24

OBOE domain model

Annotation interface

Query 1 Query 2 … ….

Online query engine

Offline data processing

Annotation as a bridge

Annotation 1 Annotation 2 Annotation 3 Annotation 4

File 1 File 2 File 3 File 4

Measurement type table

Observation Type table

Entity type table

Context type table

Map

25

OBOE-aware DB

Annotation as a bridge

Annotation 1 Annotation 2 Annotation 3 Annotation 4

File 1 File 2 File 3 File 4

Measurement type table (for ann1, 2, …)

Observation type table (for ann1, 2, …)

Entity type table (for ann1, 2, …)

Context type table (for ann1, 2, …)

Map (for ann1, 2, …)

26

OBOE-aware DB

Offline data processing (raw data loading)

Observation type table (for ann1, 2, …)

Entity type table (for ann1, 2, …)

Context type table (for ann1, 2, …)

Map (for ann1, 2, …)

Measurement type table (for ann1, 2, …)

Raw data loading

Annotation 1 Annotation 2 Annotation 3 Annotation 4

File 1 File 2 File 3 File 4

Table 4

Table 3

Table 2

Table 1

27

OBOE-aware DB

Offline data processing (data materialization)

Measurementtable

Observation table

Entity table

Context table

Measurement type table (for ann1, 2, …)

Observation type table (for ann1, 2, …)

Entity type table (for ann1, 2, …)

Context type table (for ann1, 2, …)

Map (for ann1, 2, …)

Table 4

Table 3

Table 2

Table 1

Data materialization

Annotation 1 Annotation 2 Annotation 3 Annotation 4

File 1 File 2 File 3 File 428

OBOE-aware DB

MeasurementTable (for file 1, 2, …)

Observation table (for file 1, 2, …)

Entity table (for file 1, 2, …)

Context table (for file 1, 2, …)

Measurement type table (for ann1, 2, …)

Observation type table (for ann1, 2, …)

Entity type table (for ann1, 2, …)

Context type table (for ann1, 2, …)

Map (for ann1, 2, …)

Table 4

Table 3

Table 2

Table 1

Data materialization

Annotation 1 Annotation 2 Annotation 3 Annotation 4

File 1 File 2 File 3 File 429

OBOE-aware DB

Measurement type table (for ann1, 2, …)

Observation type table (for ann1, 2, …)

Entity type table (for ann1, 2, …)

Context type table (for ann1, 2, …)

Map (for ann1, 2, …)

Table 4

Table 3

Table 2

Table 1

Query strategy 1

Online query engine (query re-writing over raw data)

30

MeasurementTable (for file 1, 2, …)

Observation table (for file 1, 2, …)

Entity table (for file 1, 2, …)

Context table (for file 1, 2, …)

OBOE-aware DB

Measurement type table (for ann1, 2, …)

Observation type table (for ann1, 2, …)

Entity type table (for ann1, 2, …)

Context type table (for ann1, 2, …)

Map (for ann1, 2, …)

Table 4

Table 3

Table 2

Table 1

Query strategy 2

Online query engine (query re-writing over materialized data)

31

MeasurementTable (for file 1, 2, …)

Observation table (for file 1, 2, …)

Entity table (for file 1, 2, …)

Context table (for file 1, 2, …)

OBOE-aware DB

Measurement type table (for ann1, 2, …)

Observation type table (for ann1, 2, …)

Entity type table (for ann1, 2, …)

Context type table (for ann1, 2, …)

Map (for ann1, 2, …)

Table 4

Table 3

Table 2

Table 1

Query strategy n??

Online query engine

32

MeasurementTable (for file 1, 2, …)

Observation table (for file 1, 2, …)

Entity table (for file 1, 2, …)

Context table (for file 1, 2, …)

Other materialization/de-

normalization??

RDF triple store?

OBOE-aware DB

Ontology Editing Tools

• Protégé plug-in - For creating and editing OBOE-compatible ontologies- Form-based UI- Generates “low-level” OWL constraints/axioms

References[1] Shawn Bowers and Joshua S. Madin and Mark P. Schildhauer, A Conceptual

Modeling Framework for Expressing Observational Data Semantic. In ER 2008, 41-54.

[2] OpenGIS observations and measurements encoding standard (O&M): http://www.opengeospatial.org/standards/om

[3] Schematron ISO standard: http://www.schematron.com/

[4] EQ: https://www.phenoscape.org/wiki/EQ_for_character_matrices

[5] CECF: http://www.cefe.cnrs.fr/ecopar/

[6] SBC LTER: http://sbc.lternet.edu/

[7] iPlant: http://www.iplantcollaborative.org/

[8] PATO: http://obofoundry.org/wiki/index.php/PATO:Main_Page

34