Wikidata:Property proposal/TRBC

Thomson Reuters Business Classification (TRBC) edit

Originally proposed at Wikidata:Property proposal/Authority control

Done: Thomson Reuters Business Classification (P5798) (Talk and documentation)

Represents	Thomson Reuters Business Classification (Q7795893)
Data type	String
Domain	organization/company OR economic activity/product
Allowed values	\d{2,10}
Example 1	Google (Q95) → 572010 Software & IT Services (Industry) (https://permid.org/1-4294951196) 57201030 Online Services (Business Sector) (https://permid.org/1-4294952710) 5720103011 Search Engines (Economic Sector) (https://permid.org/1-4294952829)
Example 2	search engine (Q19541) → 5720103011 (https://permid.org/1-4294952829)
Example 3	BP (Q152057) → 5010203010 Other Oil & Gas Refining and Marketing (https://permid.org/1-4294951991)
Example 4	gold (Q897) → 51201060 (https://permid.org/1-4294952770)
Example 5	military uniform (Q1196123) → 5210101018 (https://permid.org/1-4294951858) "Military Clothing & Accessories" (closeMatch)
Example 6	merchant bank (Q1921438) → 5510201015 (https://permid.org/1-4294951363)
Number of IDs in source	731: 26 EconomicSector, 51 BusinessSector, 104 IndustryGroup, 234 Industry, 316 Activity (these names come from ttl)
Expected completeness	eventually complete
Robot and gadget jobs	See https://en.wikipedia.org/wiki/Thomson_Reuters_Business_Classification. Import from https://permid.org/download: "industry" eg https://permid.org/sfiles/bulkDownload/OpenPermID-bulk-industry-20180805_070415.ttl.gz (need free login)
See also	NAICS code (P3224), PermID (P3347)

Motivation edit

permid offers free downloads, and we got PermID (P3347) for companies. TRBC is the industry classification used in permid. Although TRBC entries have permid URLs, the hierarchical codes (eg 51101090, 501030) are better known by people, and used in TR company data. This proposed property is similar to depicts Iconclass notation (P1257) (applied to an artwork); as opposed to Iconclass notation (P1256) (applied to a topic): many TRBC codes are not appropriate as WD entries.

https://permid.org/terms states that the industry values of companies are CC-NC (non-commercial license). However, TRBC itself doesn't fall under such license, indeed it's available on Wikipedia (but better to import from te ttl shown above).

It's already exposed on permid URLs (see examples above), but the hierarchical 2-10 digit TRBC code is more important. This should be "Wikidata property for authority control for industries and economic activities". Vladimir Alexiev (talk) 10:07, 8 August 2018 (UTC)[reply]

Discussion edit

Comment @Vladimir Alexiev: Is this an identifier in the sense of only having 1 item match an ID, or a classification as the label suggests? If the latter it probably should be string datatype (especially as we don't seem to have a formatter URL). ArthurPSmith (talk) 17:30, 8 August 2018 (UTC)[reply]
Support David (talk) 09:09, 9 August 2018 (UTC)[reply]

@jneubert: any opinion in this? Is the classification relevant enough? -- JakobVoss (talk) 21:05, 15 August 2018 (UTC)[reply]

Weak support. TRBC is one of several "standard" classifications in this field. The classes themselves are already covered by PermID (P3347), which also defines a formatter URL and a RDF URI. So we would define a subset of Perm-ID, currently without linking possiblity. @Vladimir Alexiev: Are you in contact with TR? Perhaps they would be open to implement a redirect service - even if they don't consider the classification codes as suitable for persistent URIs. Jneubert (talk) 11:35, 16 August 2018 (UTC)[reply]

Comment Another option could be using the PermID as identifier for the property, with the classification code attached in a qualifier (which of course implies a unique item ./. ID match). That would define a PermID subset for classes/classificatin codes. That may be superior to the direct use of classification codes, which sometimes change over time. I don't know if we have similar use cases elsewhere, but I think such a property could be useful for mapping the TRBC classification to WD and to other related classifications. Jneubert (talk) 11:58, 16 August 2018 (UTC)[reply]

Hi everyone!

@ArthurPSmith: This is an ID of a TRBC classification. You can apply it to a company, in which case it will classify the company. I know there's Iconclass notation (P1256) (applied to a topic) vs depicts Iconclass notation (P1257) (applied to an artwork), the latter to be used when WD doesn't have a respective item.
Maybe we need the same (or only the latter) for TRBC, because there are many TRBC codes that I don't believe will be accepted as WD items, eg 51101090 Diversified Chemicals, 501030 Oil & Gas Related Equipment and Services, etc. Comment Should I rework the proposal in this way?
@Jneubert: "one of several standard" classifications": I don't claim it is "standard" (UN ISIC, EC NACE and NAICSare standards), but it's used in TR information products, including Permid.
I've sent TR some bugs about Permid RDF data, but haven't heard back from them.
"using the PermID as identifier for the property": I believe that the hierarchical codes (eg 51101090, 501030) are better known by people, and used in company data.

--Vladimir Alexiev (talk) 13:20, 22 August 2018 (UTC)[reply]

I get the feeling from this that it would be better used in wikidata as a classifier (your "only the latter" case above) - in which case the only thing really needing to be changed in the proposal is to go from external id to string datatype and adjust your examples to be for companies rather than concepts (the "domain" value in the proposal already agrees with this). Or make two proposals, one for the id and one for the classification? Anyway we should be clear on which one this is. ArthurPSmith (talk) 13:37, 22 August 2018 (UTC)[reply]
- Reworked proposal to apply directly to company, or to economic activity/product (when such already exists in WD), and gave more examples --Vladimir Alexiev (talk) 09:18, 23 August 2018 (UTC)[reply]
Support Ok, that works. ArthurPSmith (talk) 18:33, 23 August 2018 (UTC)[reply]

@Vladimir Alexiev, ArthurPSmith, Jneubert, ديفيد عادل وهبة خليل 2: Done Thomson Reuters Business Classification (P5798) − Pintoch (talk) 08:01, 7 September 2018 (UTC)[reply]