Metadata Quality Heuristics

Grades

For each field (property):

• A - Metadata is present, complete, and correct.
• B - Metadata is present and correct, but incomplete.
• C - There is metadata, but it is incorrect.
• D - Metadata is not present, but it is known.
• E - Metadata is not present; values are unknown.
• NA - Metadata is not present but field is not necessary for this item.

If there is a concept or individual missing completely:

• X - Item is omitted entirely (and therefore all fields are either D or E)

Lists - Sensor Network

Platform Model Hierarchy

Ensure that there is one entry for each Platform Model. Verify that each entry has all of the following properties:

• Name
• Maker
• Description

Platform List

Ensure there is one entry for each Platform, and each has the following properties:

• Name (Note that this might not be unique if there are two separate platforms of the same model)
• Serial number (This must be unique for each Platform, regardless of model)
• Lat-long location

Instrument Model Hierarchy

Verify that there is one entry for each Instrument Model, each with properties:

• Name
• Maker
• Description

Not all of these are applicable to every Instrument Model; verify that these fields contain data where necessary:

• Number of attached detectors
• Maximum number of detachable detectors
• Maximum number of logged measurements
• Min and max operating temperature (+ units)

Instruments

Ensure there is one entry for each Instrument, and all have both properties:

• Name (Again, not necessarily unique)
• Serial number (Must be unique)

Detector Models

Verify there is one entry per Detector Model:

• Name
• Maker
• Description

In addition, each Detector Model should have at least one Sensing Perspective. A Sensing Perspective should have the following properties, however it is possible that not all are known for every Sensing Perspective:

• EntityCharacteristic name (Necessary)
• Accuracy %
• Accuracy R2
• Output resolution
• Max response time (+ units)
• Min and max range of values

Detector List

As with other lists of individuals, verify that there is one entry for each Detector with the following information:

• Name (Might not be unique)
• Serial number (Must be unique)

Lists - Measurements

Entities

Review list of Entities to be sure that everything your project plans to observe for data measurements is included. Note that context is very important in differentiating Entities - for example, measuring water in a lake is different from measuring a sample of water brought back to the lab.

Characteristics

EntityCharacteristics

A unique pairs of an Entity and a Characteristic, for describing Sensing Perspectives. There will not be an EntityCharacteristic for every pairing of Entity and Characteristic, only those represented in the Sensing Perspectives.

• Entity (Exactly one per entry)
• Characteristic (Exactly one per entry)

Standards

Agents

There are two different types of thing represented as "Agents" - People and Organizations. This graph should be connected, with the placeholder Agent node at the root. Edges in this graph represent a part-of/has-part relationship.

• Name (For both People and Organizations)
• Email (Where applicable)
• Website (Where applicable)
• Some parent node in the graph - an Organization may be a part of another organization, or, if it stands alone, then its parent is the Agent placeholder at the center. Similarly, a person may belong to one or more Organizations, or, if completely unaffiliated, will connect only to the Agent placeholder.