Predicate Proposal - Wikidata ID

Value statement: associating an entity with its ID on wikidata will enable data ingestion from Wikidata and provide access to a generally useful ID. Wikidata is a widely reputed source under the CC0 license that the Golden protocol can build off of.

Entity IDs can also be used as globally unique URIs that tend to follow the pattern http://www.wikidata.org/wiki/<ID> (i.e. Q12345) or http://www.wikidata.org/wiki/<instance type>:<ID> (i.e. P569) where ID is an entity ID.

Definition: The unique Wikidata ID (item, property, or lexem) associated with this entity on Wikidata. A Wikidata ID is a number prefixed by a letter. Items, also known as Q-items, are prefixed with ‘Q’. Properties are prefixed by ‘P’. Lexemes are prefixed by ‘L’.

Golden entities tend to have a 1-1 mapping with a Wikidata ID.

Tooltip definition of the predicate: “The unique Wikidata ID associated with this entity”

Type of value: String

For enums - all possible values: N/A

# of accepted values: One

Inverse Properties and Name: N/A

Examples of proper use:
- ‘Ginkgo Bioworks’ → ‘Wikidata ID’ → ‘Q29026065’ would be correct, as the wikidata page that this ID corresponds with https://www.wikidata.org/wiki/Q29026065 is the same Ginkgo Bioworks entity as the subject of the triple.

Examples of improper use:
- ‘Artificial Intelligence’ → ‘Wikidata ID’ → ‘Q4801030’ would be incorrect, as the wikidata page that this ID corresponds with https://www.wikidata.org/wiki/Q4801030 is a journal about artificial intelligence and does not represent the topic ‘Artificial Intelligence’ like the Golden page does.

Usage in other schemas:
- Usage in wikidata: N/A
- Usage in schema.org: N/A

Citation Required?: No - the ID functions as its own citation as the URL for the Wikidata page can be formed out of the Wikidata ID. ie Q12345 corresponds to http://www.wikidata.org/wiki/Q12345

3 Likes

Estimated cardinality of this predicate: There are ~100 million items on Wikidata as of July 2022 [source]

Estimated frequency of new values or changes: Wikidata ID values rarely change, primarily when the ID for that item on WIkidata is deleted. This tends to be fairly rare on average.

@jed @jen do we have an example where a Golden entity should map to multiple Wikidata IDs? The few cases I can find seem erroneous.

I assume we should allow multiple Wikidata IDs to be assigned to a single Golden entity for flexibility in the rare instance this is necessary.

To be clear, this is to specify the definition of the Predicate that already exists in Golden (or at least golden.com) today, right?

1 Like

Correct, this is specifying the definition of the Predicate that already exists on Golden.com right now. We would bring it over to the dapp with this same definition.

I have not seen more than one valid wikidata ID per entity and I just spot-checked a few entities with multiple Wikidata IDs and all were cases where one of them was incorrect. I am inclined to only accept one Wikidata ID, unless @jed has knowledge of cases where there have been more than one.

2 Likes

Yes, I should have mentioned that before. Also 100% agree on limiting Wikidata ID to just one.

Frankly, if there could potentially be multiple Wikidata IDs for one Golden entity either:

  • Wikidata has incorrect data
  • Golden should extract a single entity into multiple entities at a more granular resolution

(At least that’s my current opinion!)

1 Like

Thats a good point that we can/should encourage topics to be created in a more granular way by limiting the wikidata ID to only one per entity. Have adjusted the proposed definition above.

Turns out the few instances where multiple Wikidata IDs were assigned seems to be a UI related bug - Golden.com likewise enforces that only one wikidata ID can be added to an entity.