Wikidata:Property proposal/Tüik number

TÜİK ID

Originally proposed at Wikidata:Property proposal/Place

Not done

Description	Identifier for populated places in Turkey, in the TÜİK database
Data type	String
Allowed values	\d+
Example 1	Bilecik Province (Q46763) → 11
Example 2	Bayraklı (Q812591) → 2056
Example 3	Çiçekli (Q8002268) → 176975
Source	http://www.tuik.gov.tr/, xlsx dump
Planned use	Wikidatification of populated places in Turkey
See also	YerelNet village ID (P2123), the same for villages
Single-value constraint	yes

Motivation

Turkish Statistical Institute (TÜİK) is the Turkish government agency responsible for producing official statistics on demographics data related to Turkey. It would be beneficial for Wikidata users to have an identifier pointing to populated places in Turkey, regarding population data. --Superyetkin (talk) 14:06, 1 May 2022 (UTC)[reply]

Discussion

@Superyetkin: this request is incomplete. Please fill out the full template and add examples. Multichill (talk) 15:31, 1 May 2022 (UTC)[reply]
- Examples are available in the source file. Thanks. --Superyetkin (talk) 15:59, 1 May 2022 (UTC)[reply]
  - @Superyetkin: If you don't bother to fill out the template, we won't bother creating the property and just close it as incomplete. Please put in some effort. Multichill (talk) 16:01, 1 May 2022 (UTC)[reply]
Oppose in the absence of examples for those of us who are not tr-N. Mahir256 (talk) 19:31, 1 May 2022 (UTC)[reply]
@Superyetkin: You should provide real examples. Examples must link to any Wikidata subject (in this case, items). The xlsx you provide doesn't have links to Wikidata items, and it is written in Turkish, a lot of us are unable to understand that language in order to vote. --Tinker Bell ★ ♥ 05:05, 2 May 2022 (UTC)[reply]

Done --Superyetkin (talk) 12:43, 3 May 2022 (UTC)[reply]

Support --Tinker Bell ★ ♥ 08:15, 4 May 2022 (UTC)[reply]
Oppose this proposal - it would require five separate identifiers, as they are only unique within each type of administrative unit. For example 2056 in the example is the identifier for Bayraklı (Q812591), but also refers to Esendere (Q1367550), Kutuören (Q6448671) and Aydoğdu (Q4830384), depending on whether it is the district of Turkey (Q1147395), municipality of Turkey (Q2460358), mahalle (Q17051044) or town municipality of Turkey (Q815324)/village in Turkey (Q1529096) with that identifier (the other type is province of Turkey (Q48336), but the numbers don't reach 2056). Also as town municipality of Turkey (Q815324) is a type of municipality of Turkey (Q2460358), some places have more than one, for example Esendere (Q1367550) has 2056 as a municipality of Turkey (Q2460358), and 15377 as a town municipality of Turkey (Q815324)/village in Turkey (Q1529096) (and it looks like the districts of a metropolitan municipality in Turkey (Q2716259) have both of these in addition to one for the district) If approved, there is also the issue of different statistics for "ilçe" and "ilçe merkezi" - we currently have combined items for many of these. Q3604202 (talk) 21:31, 4 May 2022 (UTC)[reply]
Support There are 18217 villages, 32131 neighborhoods, 1366 municipalities, 981 districts in Turkey, and the en:Turkish Statistical Institute (TUIK) announces the population data for these settlements on 1 February every year. It is very difficult to manually enter the population value of more than 50 thousand settlements into wikidata every year and repeat it every year. If we match the TUIK ID number with the Wikidata item number, a bot can easily import the population data from an excel list into wikidata. --Sadrettin (talk) 16:24, 8 May 2022 (UTC)[reply]
There would have to be a way to distinguish between different units with the same ID - for example searching for the instance of mahalle (Q17051044) with value "2" would probably find Q6653955, when it refers to Q97229878 for that unit type. If every item has the correct P31, an item could be updated from village in Turkey (Q1529096) to mahalle (Q17051044) without updating the ID (perhaps because it isn't known yet) and population would be added to the wrong item. Q3604202 (talk) 12:02, 21 May 2022 (UTC)[reply]

There are no two different characteristics of settlements in Turkey. A place is either a village or a neighborhood. There is no village named "Kabasakal" (Q6653955) in Çukurova district. There is only a neighborhood called "Kabasakal". The Turkish Statistical Institute announces the population values of the "Kabasakal" neighborhood every year. Tuik ID-Wikidata ID pairings are made by checking one by one. As in the example on the right, one-to-one control is made over the excel list.--Sadrettin (talk) 17:44, 23 May 2022 (UTC)[reply]

The status of a place can change over time, and Wikidata keeps historical statistics and identifiers, and statistics are available from when it was. It's less of an issue if current status and identifier have "preferred" rank and start dates, and former status and identifier have end dates, but that isn't guaranteed to happen, and would only be possible if the identifiers were available at the same time as the status change. Queries would be easier with separate identifiers, and another reason to create separate properties is that if links become available that use these identifiers, they would probably require a different formatter URL (P1630) for each type. Q3604202 (talk) 23:37, 31 May 2022 (UTC)[reply]

Support The concerns are irrelevant. It'll be beneficial for tr.wikipedia. There are many villages, neighborhoods and municipalities in Turkey. Therefore if we don't use it, we'll be exposed to a huge workload which discourage people to update these articles.--Sabri76 (talk) 21:35, 8 May 2022 (UTC)[reply]
Support Vikipolimer (talk) 18:52, 18 May 2022 (UTC)[reply]
Support I think it will be useful for en.wikipedia ×Elvorix^talk 17:08, 20 May 2022 (UTC)[reply]
Comment @Sabri76: Can you explain why "the concerns are irrelevant"? It seems to me that Q3604202's comment above clearly shows that this as proposed is not an external ID, since the same identifier number can refer to two or more different entities. Separate properties should be proposed for each administrative level. Since "neighborhoods" are the most numerous according to Sadrettin's statistics, we could start with that, and make this proposal "Tüik number for a neighborhood" for example. ArthurPSmith (talk) 16:18, 23 May 2022 (UTC)[reply]
Support --Jelican9 (talk) 06:43, 24 May 2022 (UTC)[reply]
Support per reason -- Maurice Flesier (talk) 16:51, 24 May 2022 (UTC)[reply]
~~Strong oppose~~ (Edited to remove this if datatype is string ArthurPSmith (talk) 20:57, 7 June 2022 (UTC)) Unless this limited to one settlement type and different id's are proposed for the others. This will not work as an external id within Wikidata if the same number can be assigned to multiple entities. ArthurPSmith (talk) 13:37, 25 May 2022 (UTC)[reply]

Each settlement has only 1 attribute and 1 number. --Sadrettin (talk) 17:39, 25 May 2022 (UTC)[reply]

@Sadrettin: That's not the issue. The issue is that two (or up to five or six) Wikidata items will have the same number, if there is only the one property. That breaks the basic logic of external identifiers to identify things. ArthurPSmith (talk) 17:55, 26 May 2022 (UTC)[reply]

Support It will be very good solution for population datas. -- Zafer (talk) 10:19, 29 May 2022 (UTC)[reply]
Support it would be very useful for the Turkish Wiki. It sipmlifies many processes. Nevmit (talk) 23:52, 4 June 2022 (UTC)[reply]
Support, It will be beneficial.--Kadı Message 18:49, 5 June 2022 (UTC)[reply]

@Superyetkin, Multichill, Mahir256, Tinker Bell, Q3604202, Sadrettin: @Sabri76, Vikipolimer, Elvorix, Jelican9: @Maurice Flesier, Zafer, Nevmit, Kadı: I have changed the datatype to string, since (A) there seems to be strong support for this to be created, but (B) it does not qualify as an external-id due to the duplication of use of the same number for different entities, and (C) there is no formatter URL here anyway. If supporters are ok with this then I would consider this to be ready for creation. If you really want this as an external-id datatype then it needs to be split into separate proposals for each level as discussed above. ArthurPSmith (talk) 15:02, 6 June 2022 (UTC)[reply]

I am okay with the string data type. Thanks. --Superyetkin (talk) 15:08, 6 June 2022 (UTC)[reply]

User:ArthurPSmith I don't get it. At any point in time one Wikidata item matches one Tüik number. Over time things change like mergers and split up so multiple id's will be on one item and multiple items will have the same id, of course all qualified with start and end time. Isn't that how a lot of external identifiers work? What I just described also applies to for example CBS municipality code (P382). How is this different? Multichill (talk) 15:20, 6 June 2022 (UTC)[reply]

@Multichill: No, one number can correspond to up to 6 WIkidata items at the same time - see the example stated above by Q3604202 "For example 2056 in the example is the identifier for Bayraklı (Q812591), but also refers to Esendere (Q1367550), Kutuören (Q6448671) and Aydoğdu (Q4830384), depending on whether it is the district of Turkey (Q1147395), municipality (Q2460358), mahalle (Q17051044) or town municipality of Turkey (Q815324)/village in Turkey (Q1529096)" - ArthurPSmith (talk) 15:36, 6 June 2022 (UTC)[reply]

I'm okay

Like Jelican9 (talk) 17:08, 6 June 2022 (UTC)[reply]

String datatype is ok, this proposal can be marked as ready. --Tinker Bell ★ ♥ 20:21, 6 June 2022 (UTC)[reply]

Strong oppose Can someone explain why this should be a single property? It appears to be a set of identifiers, the spreadsheet even separates them into separate sheets. Combining them into a single property for the sake of it and using the string type because it's no longer representing an identifier does not seem like the right way to store the data. - Nikki (talk) 11:35, 18 June 2022 (UTC)[reply]

Comment I'm tired of the bureaucratic processes of Wikidata. I will no longer comment. My only goal was to quickly add population values to the pages of 50 thousand settlements in Turkey. Wikidata lost a volunteer. Sadrettin (talk) 19:17, 4 August 2022 (UTC)[reply]
Splitting into one property per subdivision level seems the best to me. --GrandEscogriffe (talk) 18:40, 15 February 2023 (UTC)[reply]
Not done, consensus has not been reached on whether this should be a singular property or multiple properties. Please attempt to obtain consensus, then mark this as ready for creation. —Tomodachi94 (talk) 07:08, 18 March 2023 (UTC)[reply]
E4024 Wooze
Notified participants of WikiProject Turkey to attempt to obtain consensus. —Tomodachi94 (talk) 07:11, 18 March 2023 (UTC)[reply]
Oppose because this property results in ambiguity. Echoing the sentiments of some, a singular Tüik number could refer to a district, a province, a city, a mahalle, or a village depending on context. However, this is a good idea, and I am not opposed to multiple properties for each of the subdivision types listed above. —Tomodachi94 (talk) 17:47, 18 March 2023 (UTC)[reply]