New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

add Technology/Technique attribute to Science Dataset #1425

Open
proccaserra opened this Issue Nov 9, 2016 · 7 comments

Comments

Projects
None yet
4 participants
@proccaserra

as per discussion with the Bioschema group at the Rothamsted meeting.
property: Technology (syn:Technique)
description: which data acquisition techniques were used to produce "variable measured"

Use case:
to be able to find datasets by data acquisition methodology(ies) used and associated to a 'variable measured'

example1:
variable measured: molecule concentration
technology/technique: "mass spectrometry" or "nmr spectroscopy" or "colorimetry" or "immunofluorescence" ...

example2:
variable measured: depression rating
technology/technique: "Zung Scale" or "HAM-D" or "Beck Depression Inventory" ...

Note:
Review health-lifescience objects and attributes:
https://health-lifesci.schema.org/MedicalTest (o)
https://health-lifesci.schema.org/usesDevice (p)
https://health-lifesci.schema.org/MedicalDevice (o)

@danbri

This comment has been minimized.

Show comment
Hide comment
@danbri

danbri Nov 9, 2016

Contributor

I think this is a good idea. We should think about a clear property name, and a definition that clarifies that we're not attempting a full representation here for (e.g. sensor) measurements.

Contributor

danbri commented Nov 9, 2016

I think this is a good idea. We should think about a clear property name, and a definition that clarifies that we're not attempting a full representation here for (e.g. sensor) measurements.

@danbri

This comment has been minimized.

Show comment
Hide comment
@danbri

danbri Nov 9, 2016

Contributor

How about calling it "measurementTechnique" (applies to types: Dataset, DataDownload, perhaps DataCatalog for small homogenous collections, value: Text or (perhaps?) URL).

Draft:

measurementTechnique: "A description of a technique or technology used in a Dataset (or DataDownload, DataCatalog), corresponding to the method used for measuring the variable(s) which may be described with one or more variableMeasured properties. "

For example, if variable measured is: molecule concentration, measurementTechnique could be: "mass spectrometry" or "nmr spectroscopy" or "colorimetry" or "immunofluorescence". If the variableMeasured is "depression rating", the
measurementTechnique could be "Zung Scale" or "HAM-D" or "Beck Depression Inventory".

Open Issues

If we have several variableMeasured properties recorded for some given data object, we have the problem of being clear which technique goes with which variableMeasured. Since variableMeasured allows for a PropertyValue structure as its value, suggest that we allow measurementTechnique to optionally be attached at that level, so that the grouping detail is preserved.

Contributor

danbri commented Nov 9, 2016

How about calling it "measurementTechnique" (applies to types: Dataset, DataDownload, perhaps DataCatalog for small homogenous collections, value: Text or (perhaps?) URL).

Draft:

measurementTechnique: "A description of a technique or technology used in a Dataset (or DataDownload, DataCatalog), corresponding to the method used for measuring the variable(s) which may be described with one or more variableMeasured properties. "

For example, if variable measured is: molecule concentration, measurementTechnique could be: "mass spectrometry" or "nmr spectroscopy" or "colorimetry" or "immunofluorescence". If the variableMeasured is "depression rating", the
measurementTechnique could be "Zung Scale" or "HAM-D" or "Beck Depression Inventory".

Open Issues

If we have several variableMeasured properties recorded for some given data object, we have the problem of being clear which technique goes with which variableMeasured. Since variableMeasured allows for a PropertyValue structure as its value, suggest that we allow measurementTechnique to optionally be attached at that level, so that the grouping detail is preserved.

@danbri danbri self-assigned this Nov 9, 2016

@proccaserra

This comment has been minimized.

Show comment
Hide comment
@proccaserra

proccaserra Nov 9, 2016

Agreed on both, definition + reliance on propertyValue structure.

Agreed on both, definition + reliance on propertyValue structure.

danbri added a commit that referenced this issue Nov 9, 2016

danbri added a commit that referenced this issue Nov 9, 2016

@danbri

This comment has been minimized.

Show comment
Hide comment
@danbri

This comment has been minimized.

Show comment
Hide comment
@danbri

danbri Nov 10, 2016

Contributor

Talking with Natasha Noy, there's still a problem maybe. We discussed adding URL range to variableMeasured for well known codes to be cited by URL/URI identifiers.

  1. We have measurementTechnique data
  2. We have several variableMeasured to describe
  3. each has a well known URL, since we then wouldn't use PropertyValue

In this case it may be that well known URLs for variableMeasured are best represented via sameAs on the PropertyValue, even if that is a bit verbose. Not clear if these 3 conditions will be common.

Contributor

danbri commented Nov 10, 2016

Talking with Natasha Noy, there's still a problem maybe. We discussed adding URL range to variableMeasured for well known codes to be cited by URL/URI identifiers.

  1. We have measurementTechnique data
  2. We have several variableMeasured to describe
  3. each has a well known URL, since we then wouldn't use PropertyValue

In this case it may be that well known URLs for variableMeasured are best represented via sameAs on the PropertyValue, even if that is a bit verbose. Not clear if these 3 conditions will be common.

@dr-shorthair

This comment has been minimized.

Show comment
Hide comment
@dr-shorthair

dr-shorthair May 9, 2017

Jumping in late: a W3C/OGC working group has been developing a revised version of the 'Semantic Sensor Network' ontology [1]. A core vocabulary has been factored out, known as SOSA (Sensors, Observations, Samples and Actuators) which is intended to have a scope suitable for schema.org . In place of 'variableMeasured' SOSA has sosa:observedProperty, whose value is expected to be a sosa:ObservableProperty [2]

ObservableProperty is intended to name a property-type, like color, mass, temperature, which appears to match the description on https://developers.google.com/search/docs/data-types/datasets .
However, in the formal definition http://pending.webschemas.org/variableMeasured the range is given as PropertyValue which appears to imply a value like 'red', or '1 kg' or '35degC' rather than PropertyType.

So, two items of feedback:

  1. consider renaming variableMeasured to observedProperty
  2. clarify range, so that it is clear that it is the semantic property type, not the number of classifier which is the property value

[1] http://w3c.github.io/sdw/ssn/ - also includes alignments with OGC O&M, OBOE, prov-o
[2] http://w3c.github.io/sdw/ssn/#features-of-interest-and-properties

dr-shorthair commented May 9, 2017

Jumping in late: a W3C/OGC working group has been developing a revised version of the 'Semantic Sensor Network' ontology [1]. A core vocabulary has been factored out, known as SOSA (Sensors, Observations, Samples and Actuators) which is intended to have a scope suitable for schema.org . In place of 'variableMeasured' SOSA has sosa:observedProperty, whose value is expected to be a sosa:ObservableProperty [2]

ObservableProperty is intended to name a property-type, like color, mass, temperature, which appears to match the description on https://developers.google.com/search/docs/data-types/datasets .
However, in the formal definition http://pending.webschemas.org/variableMeasured the range is given as PropertyValue which appears to imply a value like 'red', or '1 kg' or '35degC' rather than PropertyType.

So, two items of feedback:

  1. consider renaming variableMeasured to observedProperty
  2. clarify range, so that it is clear that it is the semantic property type, not the number of classifier which is the property value

[1] http://w3c.github.io/sdw/ssn/ - also includes alignments with OGC O&M, OBOE, prov-o
[2] http://w3c.github.io/sdw/ssn/#features-of-interest-and-properties

@thadguidry

This comment has been minimized.

Show comment
Hide comment
@thadguidry

thadguidry May 9, 2017

@dr-shorthair PropertyValue holds both sides. The "name" and "value". Where the "name" can use the propertyId or description fields ... and the value uses the value field (along with other fields if needed like maxValue and valueReference. And we allow structure for both sides, or it can be just freeform text string. Your ObservablePropety is just the "name" side and not the "value" side, correct ? I.E. "the height of a tree" and not also its "12 feet".
http://pending.webschemas.org/PropertyValue

UPDATE: @dr-shorthair Actually, I think I just found the "value" side in SOSA, looks like Observations are made and values are stored in a Result http://w3c.github.io/sdw/ssn/#SOSAResult ??
Another question I have, if that is true, is that if your SOSAResult is equivalent to our http://pending.webschemas.org/value or closer to our http://pending.webschemas.org/QuantitativeValue ?

thadguidry commented May 9, 2017

@dr-shorthair PropertyValue holds both sides. The "name" and "value". Where the "name" can use the propertyId or description fields ... and the value uses the value field (along with other fields if needed like maxValue and valueReference. And we allow structure for both sides, or it can be just freeform text string. Your ObservablePropety is just the "name" side and not the "value" side, correct ? I.E. "the height of a tree" and not also its "12 feet".
http://pending.webschemas.org/PropertyValue

UPDATE: @dr-shorthair Actually, I think I just found the "value" side in SOSA, looks like Observations are made and values are stored in a Result http://w3c.github.io/sdw/ssn/#SOSAResult ??
Another question I have, if that is true, is that if your SOSAResult is equivalent to our http://pending.webschemas.org/value or closer to our http://pending.webschemas.org/QuantitativeValue ?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment