Wikidata talk:WikiProject elections

From Wikidata
Jump to navigation Jump to search

Properties of election items[edit]

In the election infoboxes there are a lot of parameters about the previous election: before_election, before_party, posttitle, last_election<X>, seats_before<X>, seat_change<X>, previous_mps, previous_year. All these parameters are obsolete in Wikidata. The only necessary property is follows (P155) and then we can do a query to get the information about the previous election. Also parameters like "taille_drapeau" or "Largeur de l'infobox" don't fit well the structure of Wikidata because these aren't data about the election itself but about the design of an article which varies from author to author. If no one insists I will soon remove this parameters from the list. --Pasleim (talk) 17:11, 17 November 2013 (UTC)[reply]

I think that you are right, a query should be used to fetch information about the previous election such as previous_year using the property follows (P155), but for before_election, before_party there are case when the incumbent has not been elected by the previous election ; for example for 1964 United States presidential election (Q699646) before_election=Lyndon B. Johnson but he was not elected in 1960 United States presidential election (Q699590). For last_election<X>, seats_before<X>, seat_change<X>,, from my point of view as for the other properties with <X> should not be stored in an election item but in a "candidature" item defined by a couple (election, candidate) --Dom (talk) 02:24, 19 November 2013 (UTC)[reply]
Another example where values can change between elections is when politicians switch parties (known where I live as waka-jumping). --Avenue (talk) 07:35, 19 November 2013 (UTC)[reply]

Date of election[edit]

The parameter election_date, date_élection, fecha_elección (Date of election) is not so simple. For some elections there is just one date but for other there is two rounds for example 2012 French presidential election (Q487666) or more — generaly in this case an unlimited number — for example 2013 Italian presidential election (Q3586606). Do we need to have an item for the diffrent rounds? For other, the election last more than one day, for example 2013 Rwandan parliamentary election (Q14594705). How to manage these diffrent dates of election. --Dom (talk) 06:09, 21 November 2013 (UTC)[reply]

Another example of a complex election was the 2009 Indian general election (Q1134751), which was conducted in five phases during April and May 2009, one of which extended over two days.
For elections lasting longer than a day, I think we should use start time (P580) and end time (P582) instead of point in time (P585).
Where there are multiple rounds, storing the round in a qualifier might be best, especially since some elections can have many rounds. Perhaps we could use P173 (P173) for this, or create a new property for "electoral round/stage/phase". --Avenue (talk) 08:50, 21 November 2013 (UTC)[reply]
Storing the round in a qualifier sounds good however also all statistical information have round dependant values. Hence, all of them will need a qualifier too. --Pasleim (talk) 19:46, 21 November 2013 (UTC)[reply]
From my point of view we must not use P173 (P173) for this purpose as it's defiened to respect the constraint "one of": presidential election (Q858439), general election (Q1076105), municipal election (Q152450), primary election (Q669262), election to the European Parliament (Q1128324). The idea to have the property "electoral round" (with data type: Number (not available yet)) with at least the qualifiers:
--Dom (talk) 04:33, 22 November 2013 (UTC)[reply]
I think a new property is a good idea, although I'm not sure that it should have such a restrictive definition. What should we do with multiphase elections like the Indian general election, for instance? These phases are geographically defined, and are not electoral rounds in the usual sense, but I think it would be good if the new property could handle these too. --Avenue (talk) 21:57, 23 November 2013 (UTC)[reply]
You are rigth, even if phases are not handle by any infobox as it's done when there is 2 rounds. So that's mean that a phase_name property should exist. --Dom (talk) 06:23, 27 November 2013 (UTC)[reply]

Do we need "candidature" items ?[edit]

I think that we need a type of item: candidature to be able to manage the election infobox. For each election there is many candidatures from political parties or politicians — depending of the type of election — and this leed to a kind of matrix thats would be complicated to handle in one item. I suggest to have something like that:

--Dom (talk) 17:21, 24 November 2013 (UTC)[reply]

To be complete there is the property candidate (P726) that can't handel the diffrent types of candidature: person, political party or alliance.--Dom (talk) 05:55, 27 November 2013 (UTC)[reply]

Members of parliaments[edit]

How should we provide information about the members of Parliament ? I would propose to

  • create an item "member of house XX" for all national parliamentary houses (in the US, it means only "US Senator", and "US representative". For France, it means a new item each time there is a new constitution and a new type of assembly (so perhaps a dozen items for the past two centuries). This information should always be explicitly provided to avoid confusion. It means that John Boehner (Q11702) should have both United States representative (Q13218630) and Speaker of the United States House of Representatives (Q912994).
  • use this item in position held (P39), with date qualifiers, and use of (P642) to indicate which region (or which party depending on the system) she officially represents). It should be the most precise item available, so an electoral circonscription when possible. Alternatively we could use more specific qualifiers like "reprensenting XX". But we need to have a single format. See Adrien Victor Feuchères (Q2825279) as an example.
  • Do we need to state the number of the legislative somewhere (say: 42nd United States Congress (Q4637828)) ? It seems to be inferrable from the dates when it served and when someone served ten consecutive terms, that sounds a bit cumbersome.

Does that sound ok ? --Zolo (talk) 11:03, 3 December 2013 (UTC)[reply]

Some initial thoughts:
  • Positions that are not directly elected by the citizenry (e.g. Speaker, Minister, etc) are a bit different from elected positions (MP, US senator, etc). position held (P39) is probably still appropriate, but most of the other issues here wouldn't apply, and others might be more relevant.
  • Yes, the number of the legislature seems like it could usually be derived from dates when needed.
  • I like the idea of a new property "representing" for the electoral district or party they represented, but of (P642) might be good enough. How should we handle situations where they represented different electoral districts or parties at different times? We can't attach qualifiers to qualifiers. Should we have multiple statements using position held (P39), one for each district or party they represented, with the relevant dates for each? --Avenue (talk) 15:37, 3 December 2013 (UTC)[reply]
I have made a proposal for a "representing" property at Wikidata:Property proposal/Generic. --Zolo (talk) 09:11, 5 December 2013 (UTC)[reply]
Yes, I think we have to add one person several times if they are representing several electoral districts and/or political partys.
There is a special property for electoral districts somewhere.
John Smith. Office held: Senator of Farfaraway 1986-1990. Electoral district: 7th. Representing: Royal party. From: 1986. To: 1987.
John Smith. Office held: Senator of Farfaraway 1990-1994. Electoral district: 6th. Representing: New royal party. From: 1992. To: 1994.
-- Lavallen (talk) 14:57, 5 December 2013 (UTC)[reply]

P173 (P173) - instance of (P31)[edit]

What is the best way to give the type of an election?

  • For Zolo the best way is to have:
what would be most consistent with Wikidata general logic would be:
2010 United Kingdom general election (Q215622)
instance of (P31): United Kingdom general election (Q15283424)
qualifier: follows (P155): 2005 United Kingdom general election (Q428598)
And delete the "type of election" property
It's not as a qualifier, follows (P155) applies to the item, not to its nature (it's the election who is preceeded, not the nature of it). TomT0m (talk) 22:51, 11 December 2013 (UTC)[reply]
@TomT0m:: using p155 as a qualifier may sound a bit dubious from a purely logical bot of view, but that seems like a good way to avoid ambiguity about why one item can be considered the successor of another. I was prompted to propose this based on Property talk:P155 and in particular this remark by user:Shlomo: "Let's have a look on Permian (Q76402): We can state the order as a qualifier to instance of (P31) > period (Q392928) and in this case we can add follows (P155) > Carboniferous (Q133738) and followed by (P156) > Triassic (Q47158). But we can also state the order as a qualifier to part of (P361) > Paleozoic (Q75507) (not subclass of (P279) as it is misdefined now), and in this case we only can add follows (P155), since Permian (Q76402) is the last period of Paleozoic (Q75507). In the case of geological we can actually have both type of sequences defined on Wikidata and let the users (Wikipedias) decide, which one is more useful for them.". --Zolo (talk) 07:15, 12 December 2013 (UTC)Zolo (talk) 17:41, 12 December 2013 (UTC)[reply]
@Zolo: Mmm he did not mentioned the use of qualifier on instance of (P31). I'm opposed to use it that way, as neither part of, subclass, or instance of in their traditional definitions have a sequence notion. Part of has a composition, not sequence, semantics. The better way to do this would be to use a qualifier for preceeded and succeeded by [in sequence] sequence, now I'm thinking. TomT0m (talk) 13:23, 12 December 2013 (UTC)[reply]
Thinking of it a little more, maybe we can use part of, and qualify with preceded/succeeded by. We could make consistency check if the object is a sublclass of sequence, such as the serie class. We should discuss that as a general guideline for the project. TomT0m (talk) 17:30, 12 December 2013 (UTC)[reply]
@TomT0m: I don't understand what you mean by "instance of (P31) is a subclass of election"--Dom (talk) 05:53, 13 December 2013 (UTC)[reply]
Miss word /o\. I mean the object (value) of the nature of element claim (which is a class of element) is a subclass of <election>. Sayed differently, the range of instance of (P31), if I'm not confusing myself. TomT0m (talk) 10:43, 13 December 2013 (UTC)[reply]

Election complications[edit]

Elections are complicated. Perhaps we should try to make a list of the more common complications, so that the standard data structures don't end up a mess. (Feel free to expand this list.)

  1. It is frequently the case that candidates will be a member of one party, and not have that party correspond to the one that is backing their candidacy.
  2. Some elections allow for w:Electoral fusion, where multiple parties on the ballot list the same candidate. (See for example, en:United States presidential election in New York, 2012.)
  3. Some elections have groups of candidates running, such as one for president and one for vice president, or a party list in proportional parliamentary elections.
  4. In some elections, who voted for who is important data. (In situations where the electorate is a parliament, a group of nations, an electoral college, etc.)
  5. Some elections have multiple rounds, and some use voting systems such as en:Single transferable vote.
  6. Occasionally, data is available about how votes were distributed in an area. Sometimes results from particular polling stations are available, or results from particular districts or municipalities. This is valuable data that we should include if possible.
  7. Elections sometimes have "component" elections, such as European Parliamentary elections.
  8. In some elections, voters vote for both a party list and a constituency candidate(s).
  9. Opinion polls, which are also relevant data, sometimes group together multiple parties as one option.
  10. Some elections have each voter choose to support/oppose/abstain on each candidate.

--Yair rand (talk) 14:22, 20 January 2014 (UTC)[reply]

1. That is one reason I have proposed that we should use a "represents"-property instead of "member of"-property. Another reason is that "member of" is often hard to give a source for.
6. In some areas we have items about electoral districts. (Every electoral district for parliament elections in Sweden have an item as an example, also historical electoral districts have items.) They can be split in even smaller parts if we want to. That does not mean that I propose that every polling station in every election should have an item. Only that we should allow it when we have use for it.
-- Lavallen (talk) 12:43, 21 January 2014 (UTC)[reply]
Yair rand I agree with you with you that elections are complicate. The idea to have list of "complications" is a good idea. Having an ordered list would be even a better idea, because we must have a consensus on each subject. The previous subjects of this pages show that we were not able to find a solution for any of them. I propose to have the following steps:
  • list the complications to be solved
  • a vote to choose the first complication to solve
  • discuss with the maximum of wikidata users, but more important of the other wiki projects. This would take place on this Wikidata talk page.
  • choose the best solution. This would take place on this Wikidata talk page.
  • document the solution, to make it understandable to every stakeholder. This would take place on the Politics infoboxs task force page.
  • implement the solution.
For the 2 first steps I don't where this should take place. --Dom (talk) 06:29, 1 February 2014 (UTC)[reply]
Hm, dealing with these complications one at a time could be problematic, as some solutions to some issues might contradict solutions to other issues. --Yair rand (talk) 23:48, 2 February 2014 (UTC)[reply]
You are right, but I am afraid that if we don't proceed like that nothing will be done. --Dom (talk) 06:34, 3 February 2014 (UTC)[reply]

Election-results[edit]

Instead of starting new threads in WD:PP, I try to find some ideas here.

Working with old administrative division in Sweden, I have started to find a way to describe the results of local elections. In the template sv:Mall:Mandattabell rad in for example sv:Stockholms läns landsting is there often used to describe the result each year.

In these cases, the amount of information is limited. Not even the most detailed template or source here tells who the persons are behind the results.

Take Stockholm County Council (Q3233188) and the 1938-election as a workbench-object. The result was according to Statistics Sweden is: 

Number of potential voters: "the value is there somewhere"
Number of people who are disqualified to vote, because of bankruptcy (and other things):  "the number is there somewhere" 
Number of people who voted: "the sum of the below"

                   Högern  Bondeförb  Folkpartiet  Socialdemok  Kommunister Nationalsoc Others Technical_disqualified*  thereof women
Number of votes:    20391    8037        14687       61976         3136           924      372         314
Number of seats:     13        3           6           39            -             -        -                                5

*"Technical disqualified" because the content of the voting envelope contained disallowed content. (according to the election-law)

There is no article, and therefor (today) not yet any item for the 1938-election in Stockholm County counsil. But as far as I can see, we need separate items for each election. Otherwise will it be hard to maintain these items. Such a solution would also be compatible with the items about elections we already have. 1997 United Kingdom general election (Q918503) is for example an item that already exist for one (national) parlament-election. We need properties to describe the relation between UK and Q918503. In the same way, we need to describe the relation between "the 1938-election in Stockholm county counsil" and "Stockholm county counsil". (Stockholm county counsil is a municipality who is technically separated from Stockholm county. It's the English name who can be misleading.) To some extent, we maybe already have some good properties. applies to jurisdiction (P1001) or located in the administrative territorial entity (P131) already exists. applies to part (P518) and P132 (P132) are other.

Which other do we need? I guess we already have all the datatypes we need. And what is the best way to describe the above result? -- Lavallen (talk) 17:19, 21 February 2014 (UTC)[reply]

For me, property that match the best to this need is applies to jurisdiction (P1001). --Dom (talk) 13:21, 23 February 2014 (UTC)[reply]
How about this
Item:Stockholm County Council (Q3233188)
Party:Högern
Point in time:1938
Votes cast:20391
Seats won:13
same for other parties plus
party:Total
point in time:1938
Votes cast:109837
Seats won:61
Seats won by women:5
By adding the 'Point in time' qualifier you avoid the need to have separate items for each election; they can all be saved in the item for the council. Without that qualifier you need a separate item for each election and you still need all the other qualifiers for each statement.
Filceolaire (talk) 23:02, 14 March 2014 (UTC)[reply]
Another useful qualifier would probably be
part of:1938 Swedish local elections.
This would enable a query to collect all the statements relating to that election, for different local councils, and create a table with countrywide results. Filceolaire (talk) 23:16, 14 March 2014 (UTC)[reply]
That would require a property like "Seats won by women", I'm not sure that is a good idea.
What about:
Item:Stockholm County Council (Q3233188)
Num of seats: 61
Political party: All
Point in time: 1938
Num of seats: 5
Political party: All
applies to part (P518): Women
Point in time: 1938
Num of seats: 61
Political party: Högern
Point in time: 1938
Num of votes: 109837
Political party: Högern
Point in time: 1938
Num of votes: 314
Political party: Technical_disqualified
point in time: 1938
And if any seat becomes empty:
Num of seats: 1
Political party: SD
P518: Empty seat
Point in time: 2010
-- Lavallen (talk) 07:23, 16 March 2014 (UTC)[reply]
Going back to your proposal @Filceolaire:, I have tested in Stockholm County Council (Q3233188) the 2010 local election. I am still missing some way to describe how many had the right to vote in the 2010-election. And I do not have any good option for blank vote (Q2345115) (who is described as a disambig) and disqualified votes Disqualification (Q10474256) (who is about sport). It would also be nice with the possibility to add total number of votes and seats. This since somebody can vote for and represent more than one partygroup, so the sum of all 2010-parties becomes more than 100%. That becomes very obvious in the EU-elections, where most of the mp's represents both a national party and a EU group.
We also need some way to describe what the governing majority (or minority in some cases) look like. With only two partys, it's often obvious, but local elections in Sweden can result in very strange sets of majoritys, and in a few cases, no majority can be made. The majority in this specific case is set by 4 political partys (M+KD+FP+C). The head of government (P6) (landstingsråd) is here 14 persons, representing 7 different political partys. -- Lavallen (talk) 15:39, 18 March 2014 (UTC)[reply]
I do not like this idea of putting everything in the item about the council but prefer to have a separate item for every election. It's possible that the number of seats changes without having a new election (e.g. a politician changes the party) or that only a part of the council gets elected (e.g. a politician dies). So my suggestion is to have an item for every election with properties point in time, votes cast, invalid votes, blank votes, valid votes and candidate/party with qualifier votes received and num of seats. In the item about the council we have a statement for every party with qualifier num of seats, start date and end date. --Pasleim (talk) 19:53, 18 March 2014 (UTC)[reply]
Observe that this is not an item about a "counsil", it's about a "county counsil" (comparable with Landkreis), it's the English name of the entity that can be confusing.
Blank votes are here in Sweden regarded as a subclass of "invalid votes", but since a few years, they are counted separatly.
It maybe works differently in some countries, but here in "my" municipality counsil, there is 41 seats. But since nobody can claim the seat of "Sverigedemokraterna" that seat is empty. So the counsil have 40 members, but still have 41 seats. If somebody dies or is retired, the next person on the candidate-list (the next from the same political party and electoral unit (Q192611)) replaces hir without a new election. If the lists run out of valid candidates, the seat becomes empty. People here can change political party, but they cannot de jure change to another party. They can only change to independent politician (Q327591). They can de facto change to another political party, but that is less simple to add an official source to.
I agree that it would be more simple to add and read statements, both for humans and scripts, if the statements are located to separate items for each election. -- Lavallen (talk) 11:15, 19 March 2014 (UTC)[reply]

Election Results by electoral constituency[edit]

The paragraph above describes results for a total council. Where we have results for each electoral constituency we would need an item for each constituency.
For list elections the arrangement above would apply though the 'party' qualifier might be replaced by a 'party list' qualifier which could link to an item listing every candidate on the list in that election and whether they were elected that time.
For a multi-member, single transferrable vote election I would suggest the following:

Item: --Name of Constituency--
Candidate:Joe Politician
Point in time:1990
Part of:1990 general election in Foo
Votes received (STV first round):12345
Votes received (STV second round): 13456
Votes received (STV third round): 0 --this would signify that the candidate was eliminated on the third round and his votes redistributed to the other candidates--
status:not elected
deprecated
I believe this would translate into a nice table for the election results. Having the status 'deprecated' would mean that very simple query engines that don't understand qualifiers would only see the candidates elected in the most recent election (the current office holders) as only these would be preferred.
The properties for a first past the post election would then be a simpler version of this. Filceolaire (talk) 23:50, 14 March 2014 (UTC)[reply]
Would "Votes received (STV first round)" be the same property as "Votes received", where the position (first) as a qualifier, should be understood as the "first round"? Would be a good idea, if the order of the qualifier is stable, but are they? I still see the statements jumping up and down in the pages, without any intended move of that kind. How about the qualifiers? @Lydia Pintscher (WMDE): or anybody among the developers? -- Lavallen (talk) 08:34, 16 March 2014 (UTC)[reply]
That's bugzilla:58180. I _think_ it only affects statements and not qualifiers. --Lydia Pintscher (WMDE) (talk) 11:56, 16 March 2014 (UTC)[reply]
I would have separate properties for first round, second round etc. It makes it more difficult to search for the number of votes but that is all to the good since the voting numbers for second round etc. are not comparable to votes in other systems There may be a case for using 'votes received' for the first round as that is probably the one you want to aggregate to get national totals. Filceolaire (talk) 20:35, 18 March 2014 (UTC)[reply]
Depending on the type of election, the number of rounds isn't limited, see en:Swiss_Federal_Council_election,_2010#Seat_vacated_by_Hans-Rudolf_Merz. How would you express in this case the number of invalid votes in round 5 without a qualifier? --Pasleim (talk) 21:47, 18 March 2014 (UTC)[reply]

Election results, again[edit]

So, let's say there are two types of "simple" elections:

  • Type 1: All voters have the same ballots/set of choices, votes are equal, and each of the simple options on the list has a certain number of votes associated with it. The choices of each individual voter are either anonymous, unknown to the sources, or not relevant enough to include.
  • Type 2: Each voter's choice is known and relevant. Each voter present either supports/opposes/abstains on the option or options being voted on, or each voter either supports or does not support each of the options being voted on (potentially only being allowed to support a certain number (usually one) of the options).

These are fairly simple to have a data model for. While many elections are more complex than this, they can feasibly be reduced to simpler "component" elections. For example:

Splitting up elections into as many items as necessary to achieve "simple" elections could allow for a relatively simple data model. Thoughts on this? --Yair rand (talk) 18:31, 12 May 2014 (UTC)[reply]

I agree with you we must use "simple" items, for example I think even if there is for many pages in Wikipedia for general elections (for example: 2014 Panamanian general election (Q15634315)), in Wikidata we must have different items to store the results of the presidential election, the legislatives elections, the local elections, ... . --Dom (talk) 08:16, 14 May 2014 (UTC)[reply]
Dom Tubezlob Oravrattas Sarilho1 Yair rand Siwhitehouse Louisecrow Xaris333 Amadalvarez Popperipopp Dajasj Johanricher c960657 Dipsacus fullonum Zwolfz Antagonist Nikola Tulechki WikiBunny2K1 Ranjithsiji Gnoeee Kuldeepburjbhalaike PAC2

Notified participants of WikiProject Politics infoboxes

I realize this is quite an old discussion, but I came here trying to know exactly what's the standard of Wikidata regarding this issue. I'm starting a personal project to import data from French Ministry of Interior for the legislatives. Since they are several different races, I think it would make sense to separate them into various items, although I could also import the data only for the constituency (ex.: Q2973883), since there are no articles on Wikipedia about every single race. It would be nice if there was some policy in this Wikiproject regarding the issue. I personally agree to separate them, at least when the election can be split into several smaller races, although I question if the same should be made for elections such as the French presidential election. Do you have any suggestion? - Sarilho1 (talk) 23:53, 29 July 2017 (UTC)[reply]
@Sarilho1:
For me, we should have an item for each constituency. This mean 1 item for French legislative election of 2017 in Wallis and Futuna (Q24939123) (but in this case the French overseas collectivity (Q719487) is a constituency). Even in the Wikipedia in French each constituency does not exist: Category "Élections législatives françaises de 2017". For me, the French presidential election the constituency is France, so it does not make sense to get smaller items. --Dom (talk) 05:19, 30 July 2017 (UTC)[reply]

sRGB color hex triplet (P465) in political parties[edit]

Moved to Property talk:P465 --★ → Airon 90 13:19, 19 June 2014 (UTC)[reply]

Launch of WikiProject Wikidata for research[edit]

Hi, this is to let you know that we've launched WikiProject Wikidata for research in order to stimulate a closer interaction between Wikidata and research, both on a technical and a community level. As a first activity, we are drafting a research proposal on the matter (cf. blog post). It would be great if you would see room for interaction! Thanks, --Daniel Mietchen (talk) 01:48, 9 December 2014 (UTC)[reply]

My thoughts[edit]

Moved from Wikidata talk:WikiProject Elections -- Innocent bystander (talk) 09:50, 31 May 2015 (UTC)[reply]

I think we need properties for individual contests (between candidates)

We also need a slightly different set of properties for general elections (between parties).

I suggest we pick some candidates for showcase items to work on together then use as examples. Filceolaire (talk) 21:32, 24 May 2015 (UTC)[reply]

@Väsk, Wylve, Filceolaire: We (I at least) are currently testing with 2014 Swedish general election (Q7654987) and re-election of Båstad municipal council, 2015 (Q19958381). They are both "general elections in Sweden". One is on local level and one on national. To make it more complicated for you, Swedish elections are both about parties and about indiviual names. But the parties is the main issue and the only thing I have tried to describe this far.
We have to go outside of Sweden to look at individual contests, they do not exists here, and I know almost nothing about them. -- Innocent bystander (talk) 13:33, 25 May 2015 (UTC)[reply]
I added my thoughts on the use of 'successful candidate' to the talk page for 2014 Swedish general election (Q7654987); also my thoughts on New property needed for coalitions. Filceolaire (talk) 16:57, 25 May 2015 (UTC)[reply]
For elections of individuals, an example would be 2012 South Korean presidential election (Q82241). —Wylve (talk) 13:13, 26 May 2015 (UTC)[reply]
Looks very good. What I am missing is the policial parties they represent. If they represent a political party, of course. I know Iceland have had at least one non-politican as president. From the enwiki-article I see that there is a nomination process within some of the political parties. -- Innocent bystander (talk) 15:23, 26 May 2015 (UTC)[reply]
Sometimes a 'non-party' candidate can have the support of a number of parties (usually smaller parties who couldn't hope to get a party candidate elected president) so I think we need a 'supported by' property separate from the 'member of party' property. This can also be used to list other groups who are not parties but declare their support for a candidate. Filceolaire (talk) 00:42, 3 June 2015 (UTC)[reply]
Good idea, and also political candidates can have support from political parties that do not have a candidate of their own. But I would prefer 'representing' instead of 'member of political party', since you can be an MP representing a political party without being a member of it. That is also obvious when two or more political parties cooperate in an election. In the 1970 and 1973 local elections in Gotland, C and FP cooperated under the name "Mittpartierna". The individual candidates were not members of "Mittpartierna", they only represented it. In the 1973 election of Färgelanda, one MP was elected as representing both C and GV. He could not be a member of both parties, but he could represent them both. (The rules has changed since then.) In the 2014-election in Pajala a V-member (and some of his friends) wrote his own name on a SD ballot and won a seat as representing SD but he was still a member of V. -- Innocent bystander (talk) 06:36, 3 June 2015 (UTC)[reply]

Merge with Wikidata:WikiProject Politics infoboxes?[edit]

Moved from Wikidata talk:WikiProject Elections -- Innocent bystander (talk) 09:50, 31 May 2015 (UTC)[reply]

Perhaps this should be merged with Wikidata:WikiProject Politics infoboxes, particularly the discussion area? There seems to be a lot of overlap. --Yair rand (talk) 21:04, 26 May 2015 (UTC)[reply]

There probably is, maybe we can make this a subproject to Politics infoboxes? -- Innocent bystander (talk) 04:27, 27 May 2015 (UTC)[reply]
Hm, I don't think there's really enough activity to justify a separate parent project and subproject. --Yair rand (talk) 07:54, 27 May 2015 (UTC)[reply]

Gender and elections[edit]

One of the results used in templates about Swedish elections is the gender of the MP's, see under "Könsfördelning" in sv:Båstads kommun. This often interests more people than many other things in those templates. One large problem is that the sources does not tell the names of the MP's. Neither does is tell which political parties the women represents. But it still tells the number of women-seats in the elected counsils. See "Därav kvinnor" in page 168 in the pdf above.

Does anybody have any idea of how to include this information in the items about elections? @Wylve, Väsk, Filceolaire: -- Innocent bystander (talk) 11:52, 1 June 2015 (UTC)[reply]

We want to have a table with the results of the election. this has a line for each party with the number of votes received and the number of seats won. This is achieved by having a 'participant' property listing all the parties each with qualifiers for these two numbers.
To record the number of males and females elected you can use the 'number of seats won' property as a main property (not a qualifier) with qualifiers 'applies to part:male' and 'applies to part:female'. You can also use this property with no qualifier to indicate the total number of seats won in this election (there are some cases where only part of the seats are elected each time - e.g. the US senate). You can use the same pattern to indicate the number of Sammi, Catholics, russian speakers, etc. elected. OK? Filceolaire (talk) 00:35, 3 June 2015 (UTC)[reply]
Take a look in election of Båstad municipal council, 1976 (Q20019547) where I have done some experiments. (Look at it as a sandbox for now, not a standard for how it have to look like.) Your suggestion is little of the opposite of mine, but I guess it works fine. That would also allow me to add eligible voters (P1867) 78 applies to part (P518) non-Sweden (Q34) national citizenship (Q42138).
Let me set up a new page for the 1979-election according to your description, and let us compare the solutions. -- Innocent bystander (talk) 06:51, 3 June 2015 (UTC)[reply]
@Filceolaire: Maybe I misunderstood you. I have now added number of seats in assembly (P1410):13 applies to part (P518):female (Q6581072) to the 1976-item about Båstad, Sweden. I have here used number of seats (P1342) 49 to illustrate how many seats the parliament had. (They are fewer today.) In the US senate you only elect parts of the senate in each election. In Sweden, all of the MP's are elected in one election (at least today), but seats can be empty. In Kramfors there were 2010 41 seats, but only 40 MP's were elected. This happend since SD had no valid names on their ballots. How do we describe the US senate without making it look like the rest of the seats are empty? Or in the other direction, how do we make it look like 40 seats was elected, and nobody from the earlier election stayed on hir seat? -- Innocent bystander (talk) 07:37, 3 June 2015 (UTC)[reply]
Innocent bystander I think we need to have one item about the election and another about the 'council'. The election item says how many people were elected in that election. The 'council' item says how many people have seats at different times, including the gender breakdown etc. Filceolaire (talk) 08:31, 3 June 2015 (UTC)[reply]
There is an item about the counsil here. (See whatlinkshere). One problem is that my statistics only tells how many women were elected, it does not tell how many actually stayed in that seat. (Persons who choose to leave the counsil are replaced by the next available name in the ballot-lists.) Such information is it a very hard work to find, since we have to mail every municipality to get a copy of their meeting protocols to see that information. To find it for the 1910-city counsils is in several cases impossible since I know some archives has been lost. -- Innocent bystander (talk) 08:46, 3 June 2015 (UTC)[reply]

EU-elections[edit]

In 2014 European Parliament election in Sweden (Q10491412) I have added statements in the same way as I have done in the Swedish municipalities. BUT with some exceptions. I have added part of (P361) 2014 European Parliament election (Q1376095) to describe it as a constituency in the larger EU-election. I guess that can be done in all nations. The nations who are divided in smaller constituencies, can in their turn be divided in the same way, but I guess that can be optional. I have also added candidate (P726): Moderate Party (Q110843) (part of (P361): European People's Party (Q208242) to describe M as a member of EPP). In the parties who did not get any seats, I have not added any qualifier like this, since it in most cases only would be hypothetical.

Does this make sense? And would it make sense also for other elections with constituencies? -- Innocent bystander (talk) 12:45, 5 June 2015 (UTC)[reply]

And would this use of part of (P361) make sense in other kind of alliances, like candidate (P726): Moderate Party (Q110843) (part of (P361):Alliance (Q1324668)) in the national elections in Sweden in recent years? I prefer not to add such statements directly into the items about the political parties, since they cannot be used as a global variable, This group of 4 does not cooperate in all of Sweden! -- Innocent bystander (talk) 12:51, 5 June 2015 (UTC)[reply]
After I have added 2014 European Parliament election in Northern Ireland (Q20049253), I do not feel that the use of part of (P361) works in the relation M:EPP. I think it looks like we need member of political party (P102).
2014 European Parliament election in Northern Ireland (Q20049253) is very different from other parts of UK and EU. The system with single transferable vote (Q115807) requried me to use persons instead of political parties in candidate (P726) instead of political parties. It also required me to add votes received (P1111) several times fore each candidate (P726), since the ballots are counted several times.
-- Innocent bystander (talk) 07:31, 6 June 2015 (UTC)[reply]
@Innocent bystander: Unfortunately, this really does not work well. The order of qualifiers is not a reliable way to store data, iirc. Perhaps each round could have a separate set of statements? --Yair rand (talk) 12:53, 14 June 2015 (UTC)[reply]
I see, but then we probabley need qualifiers for preffered votes etc. -- Innocent bystander (talk) 12:57, 14 June 2015 (UTC)[reply]
I'm not sure I understand what you mean. Could you please clarify? --Yair rand (talk) 13:22, 14 June 2015 (UTC)[reply]
131,163 voted for Diane Dodds as their first choise (preffered). When other candidates were removed from the procedure, their ballots were counted a second time, and some voted for Diane as a second alternative. That gave Diane 131,831 votes in the second round. The most important number here is the preffered votes. -- Innocent bystander (talk) 14:49, 14 June 2015 (UTC)[reply]
I'm not completely clear on how Swedish European Parliament elections work, but aren't voters voting for a list, rather than simply for a party? If there are lists, then the lists' contents would be something we'd want to include somehow, I think. --Yair rand (talk) 09:01, 11 June 2015 (UTC)[reply]
We are voting for lists, yes. But the lists always belong to a political party. There can be two or more competing lists for one single party, and you can also vote on empty ballots with only the name of the party. You can even add names to the lists yourself, or invent new political parties if you like. The latter is not very unusual. Donald Duck gets a lot of votes every election.
You can also mark single names in these lists. If I remember correctly Maria Bildt therefor was elected first, even if she was number two in the lists of (M).
Since empty and handwritten lists are common and still eligble, it is hard to describe it in a good way. There is statistics about every person geting a vote since 1998, and that can be used, but I doubt that it is very useful. And who shuold make the work of identifing which name corresponds with an item. It's frequent that we have two or more persons with the same name in the Riksdag, it is even more common in these lists. That is probably one reason that MP's in the 19th and early 20th century often were adressed both by their name and home village. One more recent example can be found int the English sitelink of Göran Persson (Q5626648). -- Innocent bystander (talk) 10:07, 11 June 2015 (UTC)[reply]
The name overlaps may make things more difficult...
So, voters vote for lists, but they can also vote for particular list-members in order to influence the result of the list's elected candidates? I assume we'd want to store the contents of the lists regardless of whether their candidates got elected. Storing who of which list particularly got however many votes may be more complicated. Maybe under separate items? --Yair rand (talk) 12:53, 14 June 2015 (UTC)[reply]
Marita Ulvskog personally got 176,721 votes. She was the only elected candidate in (S) who got enough votes to be personally elected. The other four elected (S)-candidates, were elected as top members of the list only. 30 different (S)-candidates got personal votes.
In (KD) both Lars Adaktusson (107,897) and Ebba Busch Thor (11,167) got enough votes to be personally elected for (KD), but KD only got one single seat in the EU. -- Innocent bystander (talk) 14:49, 14 June 2015 (UTC)[reply]

successful candidate (P991) in Parlaiment elections[edit]

RE: this. I am not so sure if adding Juncker, and only Juncker is the best way to use P991 in parlaiment elections. In Talk:Q7654987 we have discussed the use of this property, but a better place for such discussion is maybe Wikidata talk:WikiProject Politics infoboxes. -- Innocent bystander (talk) 11:55, 13 June 2015 (UTC)[reply]

@Innocent bystander: Jean-Claude Juncker (Q57661) have been added in successful candidate (P991) because Dom bot add what is in the property "sucesor" for the "cargo" in the infobox in Spanish "Ficha de elección" in the page "Elecciones al Parlamento Europeo de 2014". And it is the same in other languages, the only person in the infobox is the "Elected President of the European Commission". --Dom (talk) 09:28, 14 June 2015 (UTC)[reply]
@Dom: I understand where the bot takes the information from. But the "office contested" in an election like this is not the "President of the European Commission", it's the "Members of the European Parlaiment". And the members of the Commision is not elected by the Parlaiment alone, the Counsil is also involved.
It is more or less the same thing in the Swedish parlaiment elections. The "office contested" is not the prime minister of Sweden. The new MPs elects a Chairman for the Parliament, and (s)he chooses a Prime minister. If a majority of the MP's does not vote against the new prime minister, (s)he is free to choose hir members of the new goverment. -- Innocent bystander (talk) 10:01, 14 June 2015 (UTC)[reply]
@Innocent bystander:, I understand your point of view, but you will notice that the list of members of the European Parliament, 2014–2019 (Q16958970) is not the same as 2014 European Parliament election (Q1376095). --Dom (talk) 10:21, 14 June 2015 (UTC)[reply]
@Dom:. Yes, I know! There is 349 seats in the Swedish Riskdag, but already less than a year after the election, ~50 MPs have been temporarly or permanently been replaced. And the elected MPs are not necessarily representing the same political parties today as they did when they became MPs. The same thing happens in the EU-parliament all the time. -- Innocent bystander (talk) 10:32, 14 June 2015 (UTC)[reply]
@Innocent bystander:, Do you imagine to put the 349 members of the Parliament of Sweden (Q272930) in the infobox of 2014 Swedish general election (Q7654987)? I don't, because an infobox are just summary. The aim for wikidata for me is to be able to have wikipedia pages like this one Élection présidentielle colombienne de 2010 where most of the attributes of the template "Infobox Élection" come from Wikidata. --Dom (talk) 13:01, 14 June 2015 (UTC)[reply]
While infoboxes can be a major consideration here, it's important to keep in mind the more general goals of Wikidata, which includes more general support for Wikipedia content (and of course other Wikimedia projects, and projects outside Wikimedia altogether), and not just infobox use. Data regarding which candidates won such parliamentary elections is important data that needs to be stored somewhere on Wikidata, and the election item itself may be the best place. --Yair rand (talk) 13:18, 14 June 2015 (UTC)[reply]
@Yair rand:, of course we can imagine to make other steps. but the first step was the interwiki it has been completed, the second is the infobox and it's far to be completed, for lots of reasons. We must try to end this second step before jumping to an other. We must be realistic, and to be pragmatic the only usage of claims is the infobox. Remember that often the best is the enemy of the good. --Dom (talk) 17:42, 14 June 2015 (UTC)[reply]
@Dom, Yair rand: In talk:Q7654987 we discussed some other alternatives. winner (P1346): Juncker/Juncker-commision or has immediate cause (P1478): Juncker/Juncker-commision are two alternatives. There can be other better alternatives or we can propose new ones.
Regarding adding all 349 elected MP's in the Swedish parliament election, I considered to instead put all of them in subitems about the election-results in the constituencies instead. I have already done some experiments with that in the EU-election. You can find the elected candidates in UK-Southwest, UK-Nothern Ireland and Sweden. In the EU-case, it will still be a very long list in the Germany-constituency, 96 names if I remember correctly. -- Innocent bystander (talk) 19:40, 14 June 2015 (UTC)[reply]

round[edit]

A new property proposal round has been done. --Dom (talk) 05:32, 15 June 2015 (UTC)[reply]

The proposal have been refused round and a proposition to use an item for each round as described hereafter:


--Dom (talk) 07:17, 19 May 2016 (UTC)[reply]

Wikimania 2016[edit]

Only this week left for comments: Wikidata:Wikimania 2016 (Thank you for translating this message). --Tobias1984 (talk) 12:02, 25 November 2015 (UTC)[reply]

Property proposal[edit]

Please see Wikidata:Property_proposal/Organization#Candidates_to_the_candidacy.
--- Jura 06:39, 24 March 2016 (UTC)[reply]

Two consecutive terms[edit]

Is it better to have have one statement

  • W Bush: President of the United States / 2001 - 2009 / preceded by Bill Clinton / followed by Barack Obama

or two

  • President of the United States / 2001 - 2005 / preceded by Clinton / followed by Bush
  • President of the United States / 2005 - 2009 / preceded by Bush / followed by Obama

I would prefer two, as it allows cleaner qualifiers (eg elected in (P2715)). It may make queries and reuses a bit more complex though. --Zolo (talk) 07:08, 17 July 2016 (UTC)[reply]

Property 1410 or 2124[edit]

I point out this discussion.--Caarl 95 (talk) 22:33, 11 January 2017 (UTC)[reply]

Creating WikiProject Finnish elections[edit]

We are creating a WikiProject for Finnish elections, especially the forthcoming municipal elections. I would be happy if you could keep an eye on the project and advise if necessary. Best regards, Susanna Ånäs (Susannaanas) (talk) 11:55, 3 February 2017 (UTC)[reply]

Properties of U.S. Presidential Documents[edit]

In order to address the use of Wikipedia:Template:Infobox Legislation for Executive Orders and other Presidential Documents, I've created an Wikipedia:Template:Infobox U.S. Presidential Document. Based on the properties of Executive Order 13233 (Q5419885), I've mapped certain fields in the infobox to the following properties, but want to double-check that these are appropriate before I look at populating further applicable items with Federal Register data.

Are there any properties that are more appropriate for the fields that I have described, and are there any fields that anyone can see that I have missed? — Sasuke Sarutobi (talk) 23:47, 3 March 2017 (UTC)[reply]

How about restoring WikiProject Elections?[edit]

What do you think? Could we restore WikiProject Elections and make WikiProject Finnish Elections and this project subprojects to it? – Susanna Ånäs (Susannaanas) (talk) 07:14, 5 March 2017 (UTC)[reply]

Sure, this page needs an update anyways. --
--- Jura 07:43, 5 March 2017 (UTC)[reply]

Australian politicians[edit]

Hi all,
I realise this isn't precisely the right place for this question but because there is no wikiproject "politics" or wikiproject "Australia", then this place seems the closest.
member of the Australian Senate (Q6814428) and Members of the Australian House of Representatives (Q6814382) are the two matching items for the two houses of Australian federal parliament, corresponding to the same style of [english] wikipedia article - a list of lists. However, one is classified as instance of "wikimedia list article" and the other is classified as instance of "position" subclass "senator". I'm not sure what the correct version should be, but it seems they should both be the same... Wittylama (talk) 09:34, 18 July 2017 (UTC)[reply]

@Wittylama:: member of the Australian House of Representatives (Q18912794) is the item for being a member of the House: Members of the Australian House of Representatives (Q6814382) is for the Wikipedia list of people who have been that. I've connected those up better so that this is a little more obvious. --Oravrattas (talk) 10:40, 18 July 2017 (UTC)[reply]

French Legislative Elections[edit]

I everyone, I'm sorry to bother you, but I really need help. I've already asked some questions about this here, but I still have some doubts about the format I should employ. I have already created a program in Python, using pywikibot, that allows me to change labels and descriptions and add claims to the items I want, as well as import data French Ministry of Interior website. Now it is just a question of putting it to work. Following Dom suggestion, I've started to split each department election item into each legislative constituencies election item, such as Q33998270 for Q3402037, instead of Q24938407 for Q3083, with the last turning into a parent item of the former. The problem is, election such as Q33998270 were done in two rounds. How should I state it? It seems creating one item for each round would be squeezing it too much, but there are cases were such items were created (e.g: the Presidential election - Q24102723). If they were to be created I could use the format stated in a discussion above, but if not, I'm unsure how should the statements be applied. Basically I would like to know the answers to the following:

  • Should items about elections in a given department be divided into their constituencies? (already answered, but feel free to give your input)
  • Should items about elections in a department with only one constituency have a child item, be renamed or stay the same? (e.g: French legislative election of 2017 in Wallis and Futuna constituency part of Q24939123, but there is only one Q2973883 constituency in Q35555. In theory they should be different items, but pragmatically should they be?)
  • Should two-round elections have one item each (even if the item doesn't exist in Wikipedia)? If not, which format should be adopted for the different statements?
  • Also, what would be an appropriate description for those items (I feel like the label is already a description).

Sarilho1 (talk) 14:00, 30 July 2017 (UTC)[reply]

@Sarilho1:
Here is my point of view:
  • Should items about elections in a given department be divided into their constituencies? Yes
  • Should items about elections in a department with only one constituency have a child item, be renamed or stay the same? No need to create an other item.
  • Should two-round elections have one item each (even if the item doesn't exist in Wikipedia)? yes, see my point of view in tour
  • What would be an appropriate description for those items (I feel like the label is already a description). I agree
--Dom (talk) 18:14, 30 July 2017 (UTC)[reply]

Closed list proportional representation elections[edit]

Dom Tubezlob Oravrattas Sarilho1 Yair rand Siwhitehouse Louisecrow Xaris333 Amadalvarez Popperipopp Dajasj Johanricher c960657 Dipsacus fullonum Zwolfz Antagonist Nikola Tulechki WikiBunny2K1 Ranjithsiji Gnoeee Kuldeepburjbhalaike PAC2

Notified participants of WikiProject Politics infoboxes

  • 1) Is there a standard way to store the elected candidates in a closed list proportional representation election?
  • 4) In which place should the number of votes collected by a particular list be stored?
  1. as qualifier of the list in a candidate (P726) statement?
  2. as qualifier of the list in a successful candidate (P991) statement?
  3. as qualifier of the list in both the candidate (P726) and successful candidate (P991) statements?

In my view (in this case of closed list proportional representation items) the process of adding data would be more straightforward (and potentially more easily re-usable by infoboxes too) if all the information regarding the "lists" (votes received, seats...) was managed through the qualifiers of the statements stored in candidate (P726), leaving successful candidate (P991) for the individuals. But in that case, an additional property (which I do not think it currently exists) different of quantity (P1114), should be needed as qualifier ("number of seats obtained") ¿?.

EDIT: Sorry, there is apparently a property that may be used to add the number of seats (number of seats (P1342))but it is not allowed as qualifier of candidate (P726) ?!

--Asqueladd (talk) 02:13, 19 November 2018 (UTC)[reply]

  • @Asqueladd: do you have an example of data you'd like to add (e.g. from an existing Wikipedia infobox)? This might be easier to think about with some real data. --Oravrattas (talk) 10:40, 20 November 2018 (UTC)[reply]
  • I'd recommend creating items for each party list in an election, and using those as candidates, as those are the options that the voters have to choose from. I'm unsure of how to indicate number of seats won. I don't think statement is subject of (P805) as qualifier is a correct use. How to structure election data has been discussed in general many times on this page and elsewhere, but there is no standard yet afaik. --Yair rand (talk) 21:10, 20 November 2018 (UTC)[reply]

Election - constituency relation[edit]

How should the item for a particular election within a constituency be connected to the constituency itself? One option could be to extend the use of applies to jurisdiction (P1001), but that could be confusing. Another option is to extend the use of electoral district (P768) to have it either as a qualifier for office contested (P541) or as a main value property. Or perhaps, a new property could be created. Thoughts? --Yair rand (talk) 03:29, 3 December 2018 (UTC)[reply]

The standard way of doing this seems to be with a electoral district (P768) qualifier on the office contested (P541), e.g. 2017 Stoke-on-Trent Central by-election (Q28408404) has
office contested
Normal rank Member of Parliament edit
parliamentary term 56th Parliament of the United Kingdom
electoral district Stoke-on-Trent Central
0 references
add reference


add value
--Oravrattas (talk) 06:37, 3 December 2018 (UTC)[reply]
That is outside the current scope of P768. Should the scope of the property be extended? --Yair rand (talk) 06:49, 3 December 2018 (UTC)[reply]
Adding qualifiers to office contested (P541) is fairly well-established now, so it would seem slightly odd to not be able to use electoral district (P768) as one of those. There are about 150 examples of P541/P768 pairs at the minute, across multiple countries, which is large enough to show a pattern of usage, though still small enough to make it relatively simple to migrate to some other approach, if we were to find something more suitable. This does seem like a fairly natural approach to me, though, so I'd be in favour of extending the scope accordingly and documenting this as the preferred approach. --Oravrattas (talk) 07:17, 3 December 2018 (UTC)[reply]
Seems reasonable. Started a proposal at Property talk:P768. --Yair rand (talk) 21:27, 3 December 2018 (UTC)[reply]

Electoral systems[edit]

How should data on the electoral system (eg first-past-the-post, party list, etc) of an election be structured? Should it be included on each election item individually, or for each broader type of election? Should we create a new property for this, or use something like instance of (P31), subclass of (P279), has characteristic (P1552), or something else entirely?

I also think we need a way to store data that an election is a "composite" election, that it's a group of sub-elections in different districts or a sequence at different times or some such. Ideas? --Yair rand (talk) 03:30, 3 December 2018 (UTC)[reply]

On the second half of this, there are several different forms of composites. One case is where the elections are all of different types — e.g. something like 2018 Brazilian general election (Q28220142) which has part(s) (P527) Brazilian presidential election, 2018 (Q30753034) and 2018 Brazilian legislative election (Q56351812). This is probably the most common so far, and is generally fairly straightforward, though sorting out all the links is often a little time-consuming as often the local language Wikipedia(s) of the country in question will have separate pages for each of these, but others will combine them both on to a single page (which will sometimes be named after only one of them).
For the district-by-district case, I'm guessing you mean things like 2016 United States Senate election in Florida (Q16156450), 2018 United States Senate election in Florida (Q28220897) etc, where all the sub-elections are to the same position. In general, those seem a bit of a mess at the minute. Where they get instance of (P31) claims at all, there's often no distinction between the individual elections and the group of them (e.g. ). It would definitely be good to resolve this, but I'm not sure whether that would need a slightly different approach from the Brazilian case above.
One of the things that makes this all a bit confusing is that different countries not only use different terminology for the same thing, but sometimes use the same name for different things, so the key 'base' classes have changed names a few times now. "General election", in particular, means slightly different things in (roughly) parliamentary vs presidential systems. There seem to have been a few attempts to reflect this with different Wikidata items historically, but they tend to end up getting renamed or merged. It would definitely be good to find a way to make this all a lot clearer. --Oravrattas (talk) 07:04, 3 December 2018 (UTC)[reply]

How to model primaries?[edit]

In primary elections, the positions contested (office contested (P541)) are often things like "X party's candidate for position Y in election Z", or "position on the party list for election W". What would be a good way to model this? --Yair rand (talk) 23:48, 5 December 2018 (UTC)[reply]

I can't find a lot of prior art on this one: most of the existing items for primary elections are for the US presidential primaries, e.g. 2008 Republican Party presidential primaries (Q11700469), where the office contested (P541) is to a bare President of the United States (Q11696), which is clearly insufficient. 2017 Five Star Movement primary election (Q39060822) is a little bit better:
office contested
Normal rank candidate edit
of Prime Minister of Italy
0 references
add reference


add value
But that seems back to front to me, and I suspect it might be better to go with something more like:
office contested
Normal rank Prime Minister of Italy edit
subject has role candidate
0 references
add reference


add value
I haven't given a great deal of thought to primaries so far, though, so there may be more suitable properties for this, or better approaches altogether. --Oravrattas (talk) 11:47, 6 December 2018 (UTC)[reply]
Neither of these options include the data that the election is for the particular party's candidate for the office, which is a major aspect. I wonder if some property could qualify candidate (P726) to point to the relevant primary? --Yair rand (talk) 04:19, 17 December 2018 (UTC)[reply]

Rename as WikiProject:Elections?[edit]

This current project is a result of merging WikiProject Politics infoboxes and WikiProject Elections into one, back in 2015. There has been talk a few times about splitting them again, but there doesn't seem have been enough interest, or, at least, no-one has actually done so. If we were to split them, it doesn't look like very much content would remain in the more general "infoboxes" version, as the vast majority of the discussion and documentation here has been about elections, rather than other types of political data. Does anyone object to renaming this project to WikiProject Elections? That would make it easier for people to find, and to see more clearly what the primary scope is. We could easily spin out a more general project again later if there is need for it. --Oravrattas (talk) 11:54, 20 December 2018 (UTC)[reply]

Dom Tubezlob Oravrattas Sarilho1 Yair rand Siwhitehouse Louisecrow Xaris333 Amadalvarez Popperipopp Dajasj Johanricher c960657 Dipsacus fullonum Zwolfz Antagonist Nikola Tulechki WikiBunny2K1 Ranjithsiji Gnoeee Kuldeepburjbhalaike PAC2

Notified participants of WikiProject Politics infoboxes

Is anyone working on the European elections 2019?[edit]

Title says it all... I am asking because I would like to add a "personal data" component to the European elections. If parties, candidates and elections/districts are mapped out, one can add contact information to enable individuals to ask for their personal data through Subject Access Requests. Pdehaye (talk) 12:56, 9 January 2019 (UTC)[reply]

Number of seats for election to part of parliament[edit]

I have created Q63890553 for the Greenlandic part of 2019 Danish general election (Q20495840). Two seats out of 179 seats total in Folketing (Q209151) are elected in the Greenland election. I cannot find a way to indicate that without violating property constraints. Any suggestions how to indicate that number of seats elected is 2? Note that 2019 Danish general election (Q20495840) in Greenland (Q223), Faroe Islands (Q4628) and Denmark (Q35) are conducted separately after somewhat different rules. --Dipsacus fullonum (talk) 21:33, 17 May 2019 (UTC)[reply]

New infobox full wikidatapowered and ontology proposal[edit]

Dom Tubezlob Oravrattas Sarilho1 Yair rand Siwhitehouse Louisecrow Xaris333 Amadalvarez Popperipopp Dajasj Johanricher c960657 Dipsacus fullonum Zwolfz Antagonist Nikola Tulechki WikiBunny2K1 Ranjithsiji Gnoeee Kuldeepburjbhalaike PAC2

Notified participants of WikiProject elections Hi, I am new to this project. I share with you the cawiki {{Infobox election}} now copied on the WD platform to test it easily for all of us.

The design is a combination of good practices collected from frwiki, eswiki and an old cawiki prototype with manual parameters. I supplemented this with some lessons learned working with multilingual infoboxes, such as using icons instead of text to avoid problems with different string lengths, and so on. It runs in the language of your preferences (may be with some limitations such as formats in date and other localisation aspects). As for the ontology, I follow most of the work already done on this project and the actual structures I found in the few items I found fully loaded. The most important thing for me is your opinion, discussion and, if agreed, approval of the proposed structure, because infoboxes may be different, but the data structure must be agreement for all items on the same topic. The {{Infobox election}} documentation page describes the type of elections handled, the inventory of properties managed by the infobox (not necessarily all existing for elections) and links to a specific page by each election type with its data model and one example and its statements and how to fill it out. I appreciate your comments and suggestions for improvement. Thanks, Amadalvarez (talk) 21:11, 25 August 2020 (UTC)[reply]

@Amadalvarez: This is an amazing piece of work! As I'm sure you're aware, lots of items for elections are still quite messy, as the scope of election articles often differs across the Wikipedias, with some languages tending to prefer separate items for different types of elections happening at the same time (e.g. presidential and legislative) and others tending to combine them all into a single page. These haven't always been carefully distinguished in the Wikidata sitelinks, so items have often been populated by bots drawing on data from inconsistent sources. I've been slowly tidying lots of these up, but there's still a long way to go, particularly when it comes to regional or local elections. Hopefully being able to generate a useful visual representation like this will help greatly!
It's going to take me a while to go through your model in detail, and I suspect each point will need a conversation that could get confusing quickly if they're all interleaved in one giant thread, so I'll raise most of my questions in separate sections on the Talk page for the infobox itself, rather than here. But I'd also really love for us to start updating the documentation here on the Elections project itself as well, largely based on this. The bulk of the documentation listed on this project is from a long time ago now, and was mostly copied from Wikipedia templates, rather than being "wikidata-native" or reflecting actual practice. My suspicion is that the succinct tabular version you've chosen will work very well for the as a quick-reference guide for elections in general, and we should probably just point at that, rather than having to maintain it in two different places (especially as it will be important to ensure they're kept in sync). Doing so would then also give us the opportunity to have a more verbose version here, spelling out the preferred approaches for each type of election more separately, with lots of examples etc. --Oravrattas (talk) 05:33, 26 August 2020 (UTC)[reply]
@Oravrattas: Thank you for your words. I am very encouraged.
Certainly, my idea was to use this documentation as a working basis for the overall project, and move part of it to the election project. For this reason I differentiated between a single block that explains “what the infobox expects to work” and additionally a block with the ontology for each type of election.
On the issue with WPs, it is common to many topics (Sports, organizations,..). Let’s not forget that WD was loaded from WP and this has caused a perversion to the structures that are slowly being corrected. But most WP editors have a hard time understanding that the data model is not parallel to the articles. They think that the ratio is 1:1, that is, 1 article = 1 item, 1 property = 1 line in the infobox, and so on. My infoboxes often use 2 or 3 items and each infotable line can consist of +1 property or statements that handle 7 or 8 qualifiers. For example, speaking of politicians, the P39 treatment of the cawiki infobox person manages up to 23 different qualifiers. My strategy has been to build more powerful solutions than the manual infoboxes provided and start making a gradually change. When WP editors start to see that it’s better, it’s just a matter of explaining to them “how to edit WD”. For this reason I have tried this methodology to document. I invite you to visit our showroom of WD powered and multilingual infoboxes.
As soon as we agree on a stable ontology, we can finish fixing the data that are wrong. Count on my help. I’ve started by uploading all the elections that have cawiki article, in order to encourage local publishers and have reference models nearby.
As you know this project, I suggest you copy here the doc pages you consider, then we all work on it here and, finally, I'll adapt the infobox doc pointing to the final ontology pages in the project. So, we may discuss directly in the doc pages that were residents inside the project, gathering the evolution here, because the infobox doc should not be the "official ontology place", Do you agree ?. By the way, thanks for your help with P210 affaire. It has been very helpful. Salut ! Amadalvarez (talk) 09:00, 26 August 2020 (UTC)[reply]
@Amadalvarez: Great job. I did not had the time to discover all the work that you have done. Dom (talk)

Two new properties proposal[edit]

Dom Tubezlob Oravrattas Sarilho1 Yair rand Siwhitehouse Louisecrow Xaris333 Amadalvarez Popperipopp Dajasj Johanricher c960657 Dipsacus fullonum Zwolfz Antagonist Nikola Tulechki WikiBunny2K1 Ranjithsiji Gnoeee Kuldeepburjbhalaike PAC2

Notified participants of WikiProject elections. Hi, I proposed two new properties oriented to referendums and voting for approval laws, government or parties proposal, etc. in chambers and other organizations. I talk about number of support votes and number of negative votes. It is, the typical yes and no together with abstention already handle with number of abstentions (P5043). I appreciate your suggestion in their approval pages. Thanks, Amadalvarez (talk) 12:13, 12 September 2020 (UTC)[reply]

Elections lasting several days (revisited)[edit]

I noticed this was discussed ~7 years ago, but it's unclear what the consensus is regarding the issue of elections lasting several days. According to enwiki, 2020 Egyptian parliamentary election (Q1674093) is held on 24-25 October, and a second portion of the election is held on 7-8 November. Has anything like this been modelled properly on Wikidata? Popperipopp (talk) 10:50, 14 September 2020 (UTC)[reply]

@Popperipopp: A single-round election held on multiple days usually gets more than one point in time (P585) statement. A multi-round election will usually get distinct items for each round of voting (as each will have different numbers of votes in total, of for each candidate etc, and usually also different candidates in each round.) So here I think we would want separate items for each round; each of which will then have multiple point in time (P585) statements. --Oravrattas (talk) 05:32, 5 October 2020 (UTC)[reply]

New Zealand elections, modelling electorate and list candidates[edit]

I have created Wikidata entries for all registered candidates in the upcoming New Zealand general election (as per Electoral Commission official list). However we have an unusual (but not unique) electoral system and I'm not quite sure how to model parts of it.

There are three types of candidate in the election:

  1. candidates who are standing in an electorate
  2. candidates who are on a party list (which comes with a ranking)
  3. candidates who are on both a party list and standing in an electorate.

Currently I have P3602 statements (candidacy in an election), with P706 (electoral district) statements for groups 1 and 3. However I want to find a way of showing that group #2 are not just group 1 candidates with a missing electorate, and also to allow differentiation between group 1 and 3. Since I have the ranking on the party list I want to include that too. I'm thinking of something like this example but it does lead to having an electoral district that is 'list candidate'. How does this compare to how these sorts of candidates are modelled in similar electoral systems? To be clear, party lists in NZ have no relationship to geographical locations, group 2 names never appear on a ballot paper, but inclusion on the list can result in becoming an MP. DrThneed (talk) 06:14, 5 October 2020 (UTC)[reply]

I would say that it is problem using 'list candidate' as a value for electoral district (P768). How about doing it this way:
Ryl Jensen (Q99956534) candidacy in election (P3602) 2020 New Zealand general election (Q48771246)
Qualifier electoral district (P768) <no value>
Qualifier represents (P1268) New Zealand National Party (Q204716)
Qualifier ranking (P1352) 74
--Dipsacus fullonum (talk) 15:22, 5 October 2020 (UTC)[reply]
For list candidates, isn't electoral district (P768) = New Zealand (Q664)? --- Jura 16:05, 5 October 2020 (UTC)[reply]

I don't know if I'd want to use New Zealand per se, that could be inferred by some people as meaning that there was one overall list for the country, which there isn't (the lists are created by each registered party, of which this year there are 17). "List candidate" or "party list" reflects most closely how these candidates get referred to. I've had some discussion with other Wikidata editors of parliamentary data on social media, which has suggested that we need a structure that allows for people to be on a list for one party, and standing in an electorate for a different party (we do have this situation this year) and also to allow for people to stand in more than one electorate (we don't know if that was ever allowed but it will leave room for the possibility). "List candidate" or similar is suggested as it avoids creating multiple list items (one for each party) but also is most similar to how countries like Scotland model their data, I think. So a group 3 candidate ends up with two P3602 statements, one for their list ranking and one for their electorate. DrThneed (talk) 22:44, 5 October 2020 (UTC)[reply]

electoral district (P768) defines the geographic scope .. if that is NZ, I don't see what would be wrong about it.
Also candidacy in election (P3602) could hold multiple statements for the same election. --- Jura 23:23, 5 October 2020 (UTC)[reply]
I'm not saying it is wrong, just that I think it could be misinterpreted. Anyway, it amounts to a similar thing, doesn't it, as I'd need to create an electoral district instance of "New Zealand", and we're just differing on what that item is called? Or did I misunderstand you?
Yes it can hold multiple statements for the same election but how do I enter two different P102s in that same statement to allow for a person having two different political parties, one for electorate and one for list, and make it clear which party statement applies to which (list or electorate)? That's why separate P3602s were suggested. DrThneed (talk) 23:48, 5 October 2020 (UTC)[reply]
I think it just needs this. We currently lack a property for electoral lists .. still, if a person is on two separate party lists, I think there would also be two statements. --- Jura 23:57, 5 October 2020 (UTC)[reply]
No, I still think that is misleading and could cause confusion in future years. It makes it look like a candidate is standing in a national electorate (which doesn't exist). I'm going to stick with what I have, list candidate.
You can't (I think) be on two party lists, but as per group 3 above, you can be on a party list and in an electorate but for different parties, which require different qualifiers, hence two P3602s. DrThneed (talk) 00:22, 6 October 2020 (UTC)[reply]
@DrThneed: Slightly unrelated to the main question, but worth noting that a ranking (P1352) qualifier on a candidacy in election (P3602) statement usually reflects the position in the results, not in an electoral list, for example:
  • SELECT ?item ?itemLabel ?ranking ?votes
WHERE {
  ?item p:P3602 [ ps:P3602 wd:Q63989013 ; pq:P1352 ?ranking ; pq:P1111 ?votes ]
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
ORDER BY ?ranking
    Try it!
--Oravrattas (talk) 03:41, 6 October 2020 (UTC)[reply]
@Oravrattas: Ah. Did not know, thank you. There won't ever be a result ranking for these candidates (or a number of votes) - so do you think it's alright to use it for the list ranking instead? It is the list ranking (combined with the overall votes for that party) that will determine whether they get into Parliament or not. It seems a shame not to record it, but maybe there is a better way? DrThneed (talk) 06:32, 6 October 2020 (UTC)[reply]
@DrThneed: I agree that it would be better to record this than not, but overloading this property to mean something entirely different in this sort of case is going to make it much harder to query or understand the dataset. I suspect we may want a new property specifically for the position order in a party list, but perhaps there's a different property that would be usable until then. series ordinal (P1545) isn't ideal either, and still carries a degree of ambiguity, but seems at least slightly better than ranking (P1352). Another alternative would be to explicitly model the membership of the Party List itself via a separate member of (P463) statement, where a ranking would be more meaningful. --Oravrattas (talk) 07:44, 6 October 2020 (UTC)[reply]
@Oravrattas: Thanks. It seems series ordinal is also used by some people for results ranking w.wiki/fT7. So I don't think there is a perfect solution, I may pick one for now but with the aim of coming up with a better way in future (which can be retrofitted). There's three years until the next general election at least! DrThneed (talk) 21:25, 6 October 2020 (UTC)[reply]
Nobody commented on my suggestions above, and they were maybe overlooked, so I will try to explain them better.
  1. list candidate (Q99960927) isn't an electoral district, so it is wrong to use that item as a value for electoral district (P768). Instead indicate the person isn't a candidate in a specific electoral district by using the special value <no value>.
  2. In some countries it is possible for a party to have candidates in an election which aren't members of the party, or have lists of candidates that aren't representing a political party at all, but e.g. any group of citizens or in some cases a coalition of several parties. Therefore I suggested using represents (P1268) instead of member of political party (P102) as qualifier for list candidates. --Dipsacus fullonum (talk) 20:17, 6 October 2020 (UTC)[reply]
@Dipsacus fullonum: Thank you for the suggestions. Still thinking about 1. For 2, is represents (P1268) as widely used as member of political party (P102)? In New Zealand the lists are put together by the party, so member of political party (P102) seems the right one to use. It isn't saying anything about their party membership in the context of P3602, it is saying they are standing on that party's list of candidates. But represents (P1268) could work. DrThneed (talk) 21:25, 6 October 2020 (UTC)[reply]
@Dipsacus fullonum: If it re-assures you, I agree that list candidate (Q99960927) isn't a valid value for electoral district (P768). @DrThneed: Sometimes modeling things in Wikidata brings additional insight, e.g. there is a nationwide electoral district in NZ (or I didn't get the explanation above).
Maybe part of (P361) could be an option to link to list specific items, if these are too different from party lists. I also prefer "series ordinal" over "rank" as qualifier. --- Jura 08:24, 7 October 2020 (UTC)[reply]
@DrThneed: the query you provided doesn't seem to use series ordinal, only electoral district, so I'm not quite sure what it's showing. However, of the 10 elections that have the most P3602/P1545 statements (w.wiki/fXY), 7 of them appear to use it for position in a party list, and 2 (the South Australian elections) for the position in the results (which should probably be migrated to ranking (P1352)). The Finnish municipal elections appear to use it instead of candidate number (P4243), presumably because they were created before that property existed (and so should probably be migrated to use that instead). Based on what currently exists, I'd say it's safe to use series ordinal (P1545) for the party list position (and for us to document that as the preferred modelling for that). --Oravrattas (talk) 09:44, 7 October 2020 (UTC)[reply]
@Dipsacus fullonum: can you expand a little more on represents (P1268) vs member of political party (P102)? For the purposes of querying the data (and documenting how to enter it) it would be significantly better if this were standardised if possible, rather than chosen on an election by election basis. The general preference is usually to prefer a more specific property, which here seems to be P102 rather than P1268. Do you have some examples of where P102 would be particularly confusing or misleading? Where the target isn't strictly a political party, but some other form of grouping, would a relevant instance of (P31) on that item be sufficient to explain that? --Oravrattas (talk) 06:57, 8 October 2020 (UTC)[reply]
@Oravrattas: For instance in an Australian election a candidate may either be a member of National Party of Australia (Q946040) or Liberal Party of Australia (Q241149) but at the election represents Coalition (Q1065320). Similar examples with coalitions may be found in other countries too. Another example is that in Denmark at local elections groups of citizens often form candidate lists which don't represent any party as there is no organisation or party behind the list. The candidates on these candidate lists may or may not be members of political parties, and not necessarily the same party. There is a list of Danish local candidate list at en:List of regional and local political parties in Denmark. --Dipsacus fullonum (talk) 07:33, 8 October 2020 (UTC)[reply]
@Oravrattas: Besides DrThneed wrote above: "we need a structure that allows for people to be on a list for one party, and standing in an electorate for a different party (we do have this situation this year)" So it seems that in New Zealand a candidate can represent more than one party in the same election. If these candidates aren't members of both parties, it would be false to state that by using member of political party (P102). Note that P102 is about membership, and generally you don't have to be member to candidate for a party. --Dipsacus fullonum (talk) 09:31, 8 October 2020 (UTC)[reply]
As long as it's used as a qualifier, I don't see that as a problem. --- Jura 10:46, 8 October 2020 (UTC)[reply]
NB member of political party (P102) is not specifically about membership of a party. The English label did imply that previously, but this wasn't so in other languages, or in how the property was generally used, so it was made much looser a while ago. In such regards "represents" is also quite problematic: a politician in most jurisdictions is meant to represent the people, not their party (and the target of represents (P1268) is much more often a place than a party or equivalent). --Oravrattas (talk) 10:47, 8 October 2020 (UTC)[reply]
@Oravrattas: Yes, member of political party (P102) is specifically about membership of a party. The property was proposed and created as a membership property, and the labels and descriptions in the languages which I understand reflects that. I hadn't noticed that you changed the English label and description a few months ago, and I think that it was a mistake to try to redefine the property without more discussion or support. --Dipsacus fullonum (talk) 11:16, 8 October 2020 (UTC)[reply]

P3602[edit]

Hi, I was reading previous discussion and, beside the main topic in discussion, I'd like to understand when and why should be use the P3602. I red the property discussion and the proposal form, and IMHO, it generates several redundancy information. See, for instance:

Do we need the inverse of P726?. If the previous propose of the property, was rejected, what has changed? What new arguments have been made? May be are, but I can't see it. As any of you know, I am working on defining the structure of the item of an election, and have doubts where exactly would be the name of the "person" that finally will get the seat. In the direct election (president, some kind of mayors,..) are easy. However, in elections based on the "party list" it is not clear before election day, when there are hundreds of candidates who will never get a seat, as the lists have as many candidates as possible positions and there are as many lists as to candidate political parties. After election, information may be in P39 of each person, and/or in P991 of the election item. Then, what will the P3602 be used for?. Thanks, Amadalvarez (talk) 12:42, 22 October 2020 (UTC)[reply]

Might be useful for Party-list proportional representation elections. Then you can use P726 for parties, and P3602 for individuals? I would like to use that for Dutch elections. Dajasj (talk) 16:54, 21 December 2020 (UTC)[reply]

Candidacy in election - elected or not[edit]

Dom Tubezlob Oravrattas Sarilho1 Yair rand Siwhitehouse Louisecrow Xaris333 Amadalvarez Popperipopp Dajasj Johanricher c960657 Dipsacus fullonum Zwolfz Antagonist Nikola Tulechki WikiBunny2K1 Ranjithsiji Gnoeee Kuldeepburjbhalaike PAC2

Notified participants of WikiProject elections--Vojtěch Dostál (talk) 13:31, 6 October 2021 (UTC)[reply]

I added a few hundred candidacy in election (P3602) statements for Czech legislative election 2021 (Q48834983). After the election is over, I'd like to add the result to these statements (if the person was elected or not). How would you go about that? I see that some items use results (P2501) : -elect (Q1326365) as a qualifier, but that's actually discouraged by constraints set by @CamelCaseNick:. Can you think of a good alternative? Vojtěch Dostál (talk) 13:31, 6 October 2021 (UTC)[reply]

Hi
I would say that you must use successful candidate (P991). Dom (talk) 16:01, 6 October 2021 (UTC)[reply]
In my opinion P3602 is useful before elections to indicate he/she is a candidate. After elections, if he/she got the seat must be recorded in P39 of person item; if don't, the result is not important. Regarding use P991, as Dom purpose, is correct -in the election item- if the elections are person oriented. However, the legislative elections used to be a list party voting. So, the P991 in elections item would not have persons. I invite you to see "Party-oriented case" section in this work with the election items ontology. As you can see, how and where to hold person information list, remains as "open topic to be decided". --Amadalvarez (talk) 17:19, 6 October 2021 (UTC)[reply]
@Amadalvarez Thank you for your answer, I agree with what you say about usage of successful candidate (P991). As for the rest, I am not entirely convinced - Please see my P3602 statements in eg. Q3505692#P3602. Having all those qualifiers there is very useful, but we would like to be able to see if the candidate ended up being elected or not. Using only position held (P39) is strange to me for 2 reasons: 1) Being *not elected* in election is also useful information and should be explicitly stated. 2) I am not sure if a new position held (P39) statement should also state all those qualifiers or not (they are allowed in position held (P39) constraints, but isn't that data duplication?). Cant' we come up with a better solution? :) Vojtěch Dostál (talk) 06:40, 7 October 2021 (UTC)[reply]
@Vojtěch Dostál: I do not disagree with your use of P3602 and this qualifiers are good to me. Personally, I don't like to use P3602 because is a big effort to upload/create a lot of items that never again will be useful. I mean, in Czech legislative election, for instance, we have 200 names in each list (10-12 with seat, and several other without), it is, 2400-4000 entries, that finally only 200 will be elected (and P39). A large part of the "non elected" will not have any other circumstances in the future to be admissible in WD. So the effort vs efficiency is high. I believe that P3602 could be used as a complement of P39 for those "professional politician" in order to have all his/her career when don't get seat. But not for the whole lists. Regarding your specific question about which qualifier to indicate elected/not elected, we can follow a similar rule as nominee (P2453) vs award received (P166). In this case, not all awards have P166 P2453; when someone gets the award, it's informed with P166 without change the P2453 info. So, people with P2453 and NOT P166, means he/she tried but lost. The option of use P2501, to me, is redundant, because the P39 already indicates achieving of seat, and without P39 give to us the information you commented about personal trajectory. Thanks, Amadalvarez (talk) 12:06, 7 October 2021 (UTC)[reply]
@Amadalvarez To clarify - I did not create new items for politicians, I only added P3602 to people who already were on Wikidata.
But OK, I will use P39 to indicate those who got elected, and I'll likely add to it most of those qualifiers I had used in P3602. This means that some data will be duplicated but at least easily accessible for querying from both properties. Vojtěch Dostál (talk) 13:17, 7 October 2021 (UTC)[reply]

A template for elections[edit]

In the spirit of {{Generic queries for occupations}}, I've created {{Generic queries for elections}}. I'll propose to add it to {{Item documentation}}. Do you see any other generic query that could be added to the template? PAC2 (talk) 20:19, 17 May 2022 (UTC)[reply]

Data model for candidates[edit]

I'm not sure about data model candidates for elections in the Czech Republic. Regarding the legislative, communal and regional elections, there are three types of political entities involved in the election process:

  • electoral party (candidate list) - can be a political party, a coalition (of parties or independent candidates and parties), or an association of independent candidates
  • proposing party - political party proposing a candidate within the coalition
  • political affiliation (party membership)

While the latter is a clear member of political party (P102), the first two are uncertain. represents (P1268) may be suitable for both, but then it's impossible to distinguish them. Jklamo (talk) 15:46, 20 August 2022 (UTC)[reply]

Hi, @Jklamo. I'm not shure to understand your question. May be the answer is in Wikidata:WikiProject elections/Ontology proposal for Infobox election. Although it is focussed to build the infobox, it develop the ontology for differents kind of election and results level (constituency, aggregated,..).
In section Key_properties_by_election_type explain the different qualifiers by each circumstance in case of party coalitions, etc. However, the election items are oriented to "election results". So, if elections are party oriented to obtain seats, this is the value to show in the successful candidate (P991), although people vote to a person.
You may see examples of infobox, that also are sample of well filled items of each case. In addition, in 2021 Catalan parliament election (Q48838589) you have an example of parties coalition, and in 2021 Scottish Parliament general election (Q24969168), you have a multilíder for same party/coalition.
Regarding the list of elected, it shouldn't be in election item, but be build from the position held (P39) of elected people by backlink, as I explained in above section.
I hope to help you.
Keep in touch if you need. Amadalvarez (talk) 15:04, 21 August 2022 (UTC)[reply]

Elections module for Wikipedias[edit]

Hi all!

I have started to make an elections module for Wikipedias. The hope is that it can extract data about elections from Wikidata and present them in a nice and easy way in relevant Wikipedia articles. The first version is in Module:Sandbox/Dipsacus fullonum/elections and Module:Sandbox/Dipsacus fullonum/elections/configuration. For now it has only one entry function: "elections" which makes an overview table with the number the elected candidates for each political party for a series of elections. For example for an overview of all the municipal elections in L'Arboç (Q1233527) since l'Arboç results in 1979 local election in Spain (Q113732652) use {{#invoke:Sandbox/Dipsacus fullonum/elections|elections|start=Q113732652}} which will produce this table:

YearElecto­rateValid votesPartici­pationPartiesTotalData retrieved from Wikidata
CiUPSCPSUC
19792.7471.96873,0 %54211See or edit this Wikidata item

This table is more or less a copy of the table at User:Amadalvarez/election table made by User:Amadalvarez, and I hope he will be able to use this module in the Catalan Wikipedia.

The configuration submodule contains translations for used message strings and configuration variables. I have configured the table here to present the parties with the Catalan value of short name (P1813) if available and add the Catalan label as a tooltip, or else just use the Catalan label if there is no value of short name (P1813). The party texts links to the cawiki articles when they exists.

I will welcome suggestions for any improvements and more configuration options. I also intend to add more types of tables and also welcome suggestions and wishes for this. Dipsacus fullonum (talk) 07:46, 8 September 2022 (UTC)[reply]

Thanks, @Dipsacus fullonum for your help and your initiative to create a tool.
I invite a few colleagues that usually fill election articles in cawiki. I hope they help with their knowledgement.
Regarding this table:
Open topic: in infobox election, the party evolution of votes from previous election is done between "exactly" the same party. I mean, if party A was in coalition with party B in previous election, the results are not comparables and infobox doesn't show. However, some people ask to me to compare, because sometimes the party are the same and just have some kind of branding to incorpore a motto or a supporter party without a formal coalition. Now, as you can see in this case there are parties that "complement" its name and make an artificial break. Not only it produces an extra column, but also the information is not representative of the reality. There is common situation in local elections. So, I think that we may find a solution to be able to "make comparable" two or more different parties without change the reality.
I'm working a proposal and need to check that doesn't affect to normal running. Tomorrow I'll make a presentation.
Keep in touch, Amadalvarez (talk) 18:44, 8 September 2022 (UTC)[reply]
I think I can do the suggestions for headers, columns and rows. But I am not sure how to best scale logos as the same width for all logos may not be optimal. I could write the highest number of elected candidates in a row in bold, change the colors, or maybe just add a CSS class to the cell(s) so the actual formatting can be done by a templateStyle CSS file. --Dipsacus fullonum (talk) 20:57, 8 September 2022 (UTC)[reply]
Hi, it's improving. Great !!
  • CSS, it's good to me. I use it in infoboxes in order to allow personalize easily when are exported to another WP. I believe we need a class for highlight the highest (may be bold, may be a palid colour in background,..), another for the total seats column (similar to previous class: may be a grey background or bold text), another for header background & last one to body background; in these last cases is good to me as is now, but it helps to personalize.
  • half-header with "parties" text, could be delete. I made it to show in demo page. However, when the table being inside the article, it will be obvious.
  • May we have thousands separator (period or comma)?
  • Could we have a manual parameter to setting first-year last-year to show?. If we have long lists may be we wish to short them or split in two,..
  • I understand the problem with scalable logos. However, assuming the short-name usually are 5 or less digits and result figures are under 1000, could we used an equivalent width to 4 capital letters for all logos?. Only in case the name was full because short did not exist, the logo will seems small.
  • Last: I appologize, but I was wrong with the qualifier of P991 to be used. IT'S NOT quantity (P1114), but number of seats in assembly (P1410). I already change them start changing now. If you make a test meanwhile, don't be scared.
Amadalvarez (talk) 10:55, 9 September 2022 (UTC)[reply]
@Amadalvarez: For the municipal elections in l'Arboç I cannot use ballots cast (P1868) and eligible voters (P1867) to calculate for participation percentage as you suggest as there are no data for these two properties. I can use electorate (P1831) and either subtract number of abstentions (P5043) or sum total valid votes (P1697), number of spoilt votes (P5044) and number of blank votes (P5045) to get how many voted (probably best to do both when possible and warn for inconsistencies).
quantity (P1114) is still used as qualifier to office contested (P541). I suppose it is safest both to check for quantity (P1114) and number of seats in assembly (P1410). --Dipsacus fullonum (talk) 18:55, 9 September 2022 (UTC)[reply]
P1868 has not been uploaded by my fault. Tomorrow I'll do. >>> ✓ Done
I use eligible voters (P1867) or electorate (P1831), in this order. Honestly, I did not understand very well their exactly differences, but I filled P1831, in these cases.
I used quantity (P1114) for P541, because the definition of number of seats in assembly (P1410) talk specifically about "seats for a party". However, I agree to check for P1114 or P1410, just in case.
Thanks, Amadalvarez (talk) 23:38, 9 September 2022 (UTC)[reply]
@Amadalvarez, I think I have implemented most of things you mentioned. Party logos are still missing because I am waiting for your proposal for how to handle "comparable" parties which will certainly require rewriting of the code for the table headers anyway. The parties now appear in random order, but the party columns should probably be sorted in some way, I'm just not sure what to use as sort keys. Should there be a default color to use for parties without sRGB color hex triplet (P465)? I added a column with Wikidata links. I think it should be configurable which columns to have as people are not likely to agree about that. Here is a test of the new first_year and last_year arguments
YearElecto­rateValid votesPartici­pationPartiesTotalData retrieved from Wikidata
What is still missing now before it can be used? --Dipsacus fullonum (talk) 14:45, 10 September 2022 (UTC)[reply]
@Dipsacus fullonum Woow !. It looks wonderful.
  • Now I'm testing in {{Infobox election}} the "solutions proposed" to compare results between parties. I think it will be running tomorrow.
  • The default colour I use is #888. I understant it will be in the template styles. If you wish, we may use Template:Infobox election/styles.css to add the new module CSS for this and futures developements. It's up tp you.
  • to order parties, I guess it could be by the number of participations and the better total results. So, the most regulars would be at first columns, ordered by the most to less voted by the same number of appearance.
  • Wikidata links: I see now that the link of year point to the edition of the elections, but I expected the year pointed to the item where the data came from; or the article, if there is a sitelink for the corresponding "lang". Could it be ?. Then, we won't need the new column.
  • selection by year: Perfect !
  • List of entries: I figure out that you start by one item and follow the followed by (P156) to get the items of the list. Am I right?.
I selected by type of election in P31 for one applies to jurisdiction (P1001). But your way seems valid if the serie is correctly built. At this point I have to inform that the P155-P156 are -generally- as a principal property in election items. However, in this serie for "local elections in l'Arboç" I'm testing moving P155-P156 as a qualifiers of P31.
This change responds to this proposal. Now, as I may check, there already are items with serie as a qualifiers, but not so easy as my sample, because some of them have more than one P31 with serie, as 1966 United States elections (Q7892439). My idea was to migrate the main properties to a qualifiers. But I need to study more cases, to have a plan.
I think you may keep the present way and, alternatively, use the main property P156 when P31 doesn't have it. Actually, I'll use this formula during migration period.
Thanks ! Amadalvarez (talk) 19:43, 10 September 2022 (UTC)[reply]
@Amadalvarez: The module (or later template) is given the item id for the first election (Here: l'Arboç results in 1979 local election in Spain (Q113732652)) as an argument in the call, and then the code follows the chain of followed by (P156) values. In a module, it is possible to extract statement and qualifier values from any known items, but it is impossible to find new items with a known value. So I cannot e.g. find the items which have a statement applies to jurisdiction (P1001) with L'Arboç (Q1233527) as value. Just like SPARQL has limitations with limited general data processing, so has modules limitations in how data from Wikidata can be found. But it is no problem to get followed by (P156) either as a qualifier to P31 statements with a specific value or a qualifier to any P31 statements or as an statement value.
I think it will be confusing for a Wikipedia reader if the year links to the Wikidata item. In Wikipedias (at least in dawiki where I am active) ordinary blue links should never point outside of the Wikipedia. How about placing a pencil icon with the Wikidata links in the year column after the year which then still can link to the best available Wikipedia article? --Dipsacus fullonum (talk) 07:05, 11 September 2022 (UTC)[reply]
@Dipsacus fullonum Ok to the access method. I did not know this limitation you tell and I was influenced by SPARQL backlink. No problem starting in a specific item and follow P156.
If any WP doesn't link to WD from a blue link, I think the cleanest look is as you propose, it is, with a hideable link via parameter in last column. For your information, in cawiki when a blue link goes to WD, the module:wikidades changes colour link to a different tone of blue (see infobox in ca:Eleccions_al_Parlament_de_Catalunya_de_2021). It allows to know information from person, institutions, etc. knowing in addition, that WP article doesn't exist. Showing a red link, we don't know if WD exists and, to find it to see any information, we need a longer way.
Open topic: Regarding P155-P156 as a qualifiers, we need to solve how to identify what you say "qualifier to P31 statements with a specific value". As there are multivalues in P31 with its own P155-P156, we need a easy way to identify "the good one" to our objective. It affects to infobox election too and I won't start to move P155-P156 to a qualifiers until we have a criteria to choose correctly.
Thanks ! Amadalvarez (talk) 07:45, 11 September 2022 (UTC)[reply]
@Dipsacus fullonum. Finally I got the solution about "equivalent party":
  • Using said to be the same as (P460) as a qualifier of candidate (P726) with the Qid of the equivalent party as a value.
  • Its equivalence is valid just to compare with previous election, when we manage only one edition. I fill all the affected items of sample, so taking the P460 to group will be good to this module.
  • It doesn't change nor affect to the "formal structure" of item, so real name is not affected.
  • This qualifier is out of typical elections data dictionari, so there is no risc to overlap with some other use.
  • It must be a qualifier rather than a party property, because equivalence does not apply to all jurisdictions, not even to the same election edition.
  • It is in P726 because it applies from the candidature moment.
Thanks, Amadalvarez (talk) 07:10, 12 September 2022 (UTC)[reply]
I have been busy today, so I will look at it tomorrow. It will be a challenge but I think I can manage. --Dipsacus fullonum (talk) 14:07, 12 September 2022 (UTC)[reply]
Hi, @Dipsacus fullonum. Do you think you will have a chance to finish the module?
Come on, there's not much left!. Thanks, Amadalvarez (talk) 09:50, 29 September 2022 (UTC)[reply]

Wikidata Lab about elections[edit]

Hello to all. I am organizing a Wikidata Lab next week (September 22, at 17h (UTC)) about modeling elections and want to bring one or more people with experience to discuss which are the best practices to model this type of information on Wikidata and talk about their experiences and present the workflow one should follow. Are you interested in present? Please let me know by messaging me or sending me an email (eder.porto@wmnobrasil.org). If you are interested but the date is not convenient, we have some flexibility to chenge it. If you have any questions, please ping me! Cheers, EPorto (WMB) (talk) 18:53, 12 September 2022 (UTC)[reply]

Amadalvarez, Dipsacus fullonum, PAC2, Oravrattas and others ( Notified participants of WikiProject elections and WikiProject every politician has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.) EPorto (WMB) (talk) 18:57, 12 September 2022 (UTC)[reply]
Hi, @EPorto (WMB). Thanks for your invitation. I assume that you already know the document done in 2021 and currently active for {{Infobox election}}. All the structures created or used by this infobox since then follows the defined ontology.
Although the document is only defined for the representation of election results with just the data dictionary it needs, the definition of typologies according to each type of voting is valid for other uses.
From a temporal perspective, we may consider 3 moments:
Pre-electoral: candidacies, slogans, debates, polls, forecasts, lists of candidates, etc.
Elections: official results of the voting process. Without knowing or applying the calculation criteria.
Post-electoral: constitution of parliaments, attribution of office to the person (P39), election of president among candidates (indirect elections).
The election infobox provides service from the end of the pre-electoral period, when the nominations are closed (and we have uploaded them to the P726) until the beginning of the post-electoral period (distribution of seats between parties without the "official" name of the deputies, with the name of the elected president without having taken office,..). Infobox doesn't know nor apply the calculation criteria (minimun to the party get access to parliament, D'Hondt method or similar, etc.)
What is your idea about the scope of Lab ?
Thanks. Amadalvarez (talk) 05:13, 14 September 2022 (UTC)[reply]
Hello, @Amadalvarez. I saw the award, but confess I didn't go deep into it. My idea for the lab is exactly what you presented in your comment: (1) To present the standard modeling of elections and topics around it, and going towards (2) What could we do to fit the Brazilian items to this modeling. The end result is to document the process and to train whoever is watching the presentation on how to execute.
We did some modeling a few years ago (w:en:Template:Mbabel-BSE (in English) and w:pt:Wikipédia:Mbabel/Eleições brasileiras (in Portuguese)) on the topic, but it's since then cold. I was wondering how they relate and what should we do to model that information into this modeling. EPorto (WMB) (talk) 22:40, 15 September 2022 (UTC)[reply]

Ask for a doubt[edit]

I would like to extract voter turnout data from past elections (primarily national, but local elections could be interesting too) in different countries and see if there is some statistical correlation with voting system used, number of seats per constituency etc. It may be a little premature to expect any reliable results at this stage, but consider it a test case to find out what structures and properties would be required for Wikidata to support such queries. Right now I don't know how to determine which constituencies simultaneously pertain to a particular elected assembly (senate, house, unicameral legislature or state/regional/municipal/local government assembly), nor find out the number of seats assigned to each constituency. Searching for instances of electoral district (constituency) associated with a single country yields way too many of them, as they may pertain to regional/local assemblies or an old, bicameral legislature (prior to 1970 om Sweden). I deprecated a claim making all municipalities sub-classes of election district; that brought down the apparent number of constituencies in the USA from more than 30,000 to a more reasonable amount, and likewise in other countries.
Do we need a new property "pertains to elected assembly" for constituencies and elections alike? I'd prefer not to use a generic property such as instance of (P31) or part of (P361); that's likely to have ambiguous interpretations. SM5POR (talk) 23:11, 17 October 2022 (UTC)[reply]
I'm not sure to understand your question. What about to use parliamentary term (P2937) ?. As it's described in Template:Infobox_election, Operation section, it could be used as a qualifier of office contested (P541) for parlamentaries elections, it is, starting a Legislature.
However, if you wants to know the legislative term (Q15238777) that overlaps with other type of elections, we may use parliamentary term (P2937) as a main statement of the election. In any case, you know that a legislature may not fit with municipal or presidential period. So, we will have multiple values.
Another option, similar as you propose for new property, could be to use partially coincident with (P1382), instead.
Regarding number of seats assigned, the model describe that are as office contested (P541) qualifier number of seats (P1342) (or maximum capacity (P1083), when elections are just for a part of the chamber).
Kind regards, Amadalvarez (talk) 16:24, 18 October 2022 (UTC)[reply]
Thanks, parliamentary term (P2937) may well be a useful key property to look out for (assuming that it may apply to any kind of assembly, not just national parliaments or legislative bodies). But I still don't understand how to find all the constituencies involved in an election (or defined for an assembly at a particular point in time), as the data entered about Swedish constituencies is very limited, effectively confusing different assemblies over a long period of time with each other. I could probably add that information myself, once I understand the model meant to be used. Can you point to an election anywhere with all the properties defined to make it a suitable template or model item?
Also, did you really intend to refer to maximum capacity (P1083)? It seems to be about vehicles, physical venues and public events (such as maximum number of spectators allowed due to safety considerations), not formal seats in an assembly or contested in an election... SM5POR (talk) 20:41, 18 October 2022 (UTC)[reply]
I assume you know the ontology. "Item levels" section is very usefull to understand the different structures for each kind of level (from constituency to groupped information). It defines that part of (P361) link each level to the upper one. A good example are 2021 Catalan parliament election (Q48838589) and its constituencies. All cases shown in infobox examples are well filled cases for different type of elections.
Number of seats "to fill in this election" are defined by P1342. It is, usually, 100% of seats available, the full chamber is renoved. However, some chambers call for elections just of a part of its seats; in these cases the P1342 define seats for this election and P1083 the total seats of chamber. P1083 is "number of people allowed for a venue or vehicle".
By the way, I created a new section for this talk to separate from previous one. Thanks, Amadalvarez (talk) 04:45, 19 October 2022 (UTC)[reply]
Thanks a lot for those pointers; unfortunately I had overlooked the ontology you refer to. I will have to study that, as well as the election example items you mentioned.
But I maintain that maximum capacity (P1083) is not meant for governing assemblies; the expression "maximum capacity" refers to the highest number of people allowed to simultaneously enter a building or an arena, not the total number of formal seats (i.e. official positions, not physical chairs or other pieces of furniture) making up the assembly. I find it a bit of a stretch to equate an assembly with a venue, and although both terms may apply to a parliamentary building, I'd say the "maximum capacity" number would include not only the elected members of said parliament, but also their assistants, operational staff such as janitors and security, news reporters, and visitors who want to follow the debates as spectators. There is even an explicit constraint on maximum capacity (P1083) against using it for organizations, to encourage editors to create separate items for, say, corporations of performing actors and the theatrical buildings they typically perform in; i.e. you may use maximum capacity (P1083) on the latter, but not on the former. And a governing assembly is indeed an organization.
Why do you need maximum capacity (P1083) anyway? Since number of seats (P1342) is explicitly defined as "total number of seats/members in an assembly" without reference to an election, I find it counter-intuitive to effectively redefine this property to mean "number of seats contested" in a particular election. I would rather use number of seats in assembly (P1410) ("number of seats a political party, faction, or group has in a given legislature"), where the "group" might be construed to refer to seats assigned to a particular constituency. Interestingly, number of seats in assembly (P1410) has no item type constraint, which would allow us to use it directly on a constituency, a political party, or a conspiracy theory (Q159535) for that matter! En each case, the value refers to a subset of the seats, as identified by the item together with any qualifiers added. SM5POR (talk) 15:05, 19 October 2022 (UTC)[reply]
Don't worry by the use of maximum capacity (P1083). It's no longer in the infobox documentation. I rather to take it out and facilitate you focus in the important topics about elections.
Remind that qualifiers are properties of the relationship between one property and its value. It is not a property of the item, so its meaning is not necessarily conditioned by the domain of item. For instance, number of seats (P1342) is a property of the chamber. However, it appears as a redundant information to avoid a mistake, if someday the number of seats changed in chamber item.
The number of seats in assembly (P1410) are "number of seats a political party, faction, or group has in a given legislature", not the "total" as is P1342. So, I consider better to define it as qualifier in successful candidate (P991), and use P1342 for the whole seats of office refered by level election of the item. I assume you have seen the P1342 for constituency level items is the "total seats" assigned to this constituence not the chamber.
I mean, the meaning of properties is relative to the item where are used. The items for elections results are similars among them, although belong to a different level of information and type of elections with different rules. It's not necessary to have a specific property for each specific situation.
I appreciate your suggestions, Amadalvarez (talk) 15:39, 23 October 2022 (UTC)[reply]
I agree that properties should be generic enough to be useful in multiple situations, to avoid having to create multiple properties with near identical meaning. I do however see a problem when such a generalization contradicts the explicit wording used when the property was created; if you don't establish consensus in support of a broadened interpretation and have the documentation updated accordingly, you risk having different groups of editors apply diverging interpretations of the same property.
As I'm a relative newcomer in this discussion, I won't insist that one interpretation be applied above all the others, but I rather try to find out what the current consensus is and stick to that.
There is at least one qualifier that seems to contradict your claim that qualifiers don't relate to the domain of the subject item, and that is subject has role (P2868). Another is identity of subject in context (P4649), I think. This issue goes well beyond the subject of elections, however, and therefore I won't go into detail about it here (nor do I see an imminent need to discuss them right now). SM5POR (talk) 20:44, 28 October 2022 (UTC)[reply]

P155-P156 as a qualifier[edit]

Dipsacus fullonum, PAC2, Oravrattas, EPorto (WMB) and others (

Dom Tubezlob Oravrattas Sarilho1 Yair rand Siwhitehouse Louisecrow Xaris333 Amadalvarez Popperipopp Dajasj Johanricher c960657 Dipsacus fullonum Zwolfz Antagonist Nikola Tulechki WikiBunny2K1 Ranjithsiji Gnoeee Kuldeepburjbhalaike PAC2

Notified participants of WikiProject elections and WikiProject every politician has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.).

There is an open and large debate presented by @Mahir256 in "request for comment", which propose that P155/P156 should allways being qualifiers of something that, actually, is the serie or timeline that make sense to talk about previous-next. I'm in favour of this propose, and I want to know your position on it and define which migration to do. To know what we have now and decide further actions, you can consult:

cc:@Davidpar, KRLS:

I appreciate your comments. Thanks, Amadalvarez (talk) 08:11, 15 September 2022 (UTC)[reply]

P210[edit]

Hello, right now party chief representative (P210) is recommend as a qualifier in this project. However, the property does not allow this, see 2021 Dutch general election (Q42733725) for example. Should we change the property, or should we change the guidelines within this project? Dajasj (talk) 16:49, 31 October 2022 (UTC)[reply]

@Dajasj ✓ Done, see Property_talk:P210#Use_as_qualifier Amadalvarez (talk) 08:19, 3 December 2022 (UTC)[reply]
Thanks! :) Dajasj (talk) 12:22, 3 December 2022 (UTC)[reply]

P102 vs P1268[edit]

Okay another question as well. 'Questions to solve' asks: "Some candidates are not a member of the political party they represent, or they change membership after the election. How should that be described?". Right now the project recommends using member of political party (P102), which indeed assumes a formal membership (and a party, which isnt the case always in the Netherlands, you can also simply be a list). Could it be useful to use represents (P1268) instead? This avoids the formal party membership part, while keeping the same information. member of political party (P102) can still be used on the page of an individual to inform people about the membership. I'd love to hear your opinions and maybe change this in the future. Dajasj (talk) 17:06, 31 October 2022 (UTC)[reply]

@Dajasj In Wikidata:WikiProject_elections/Ontology_proposal_for_Infobox_election you will find a list of specifications created to be used by {{Infobox election}}. Now is not a proposal, but a reality, because it has been running since 2019 without collisions.
Its descriptions are oriented to explain the ontology and data dictionary for each type of elections. It's, define more specifically the properties to use, upgrading generic info described in the project main page. (when I had spare time, I'll update main page...Sorry.)
I recommend to read entire document, but in this section you will find your question answer when describes P726 qualifiers. Effectively, P1268 is better than P102, and similar to party chief representative (P210), in this circumstances.
Salut! Amadalvarez (talk) 08:41, 3 December 2022 (UTC)[reply]
Hi @Amadalvarez, thanks for your response and the infobox. It has been interesting to read and might have to adjust some things I have done for the Dutch elections (which doesnt really fit the model perfectly). Anyway, I am still confused by the 1268 vs 102, because the section mentions both as an option, as does Template:Infobox election. Template:Infobox election/basic examples doesnt use it either, and Ramon Espadaler Parcerisas (Q9066280) for example, uses 102. I am ofc willing to change the qualifier, but I do feel like P102 is more specific at least for the Netherlands, because you dont have to be formally member of a party. Dajasj (talk) 19:50, 3 December 2022 (UTC)[reply]
In party-list proportional representation (Q31764), usually for representative chambers, we vote to a party list and it has a lead candidate (Q42417559) that is the party chief representative (P210), either being party member or not.
In person-oriented elections (direct, indirect, limited multi-vot,..) we vote to a person that may be member of political party (P102) or just represents (P1268) it.
The Ramon Espadaler i Parcerisas (Q9066280) case is a mistake. Observe that in the infobox he doesn't appear as a represent of Democratic Union of Catalonia (Q1628993), because is a party-oriented election and should be P210 instead P102.
----
By the way, I encourage to use {{Infobox election}} in nlwiki. It's 100% WD powered, is multi-language. See the full example cases to know its features. I you decide to use it, I'll help to install it. Amadalvarez (talk) 21:46, 3 December 2022 (UTC)[reply]
Im not sure I understand you. Obviously the Netherlands is in the first category. But Im also adding the individual candidacies of persons. And in those cases I dont know their formal party membership, but I do know they represent the party on that list. That's why I would prefer P1268, which is in line with what you write as far as I understand.
I would love it if the Dutch Wikipedia would use the infobox, but the average Dutch Wikipedian does not agree unfortunately, but I'll keep an eye on the possibilities. Dajasj (talk) 21:51, 3 December 2022 (UTC)[reply]
When you said "the Netherlands is in the first category", do you mean that your elections are proportional party oriented?. Is it?.
What is the individual candidacies of persons you're adding?. In 2021 Dutch general election (Q42733725), I can't see it? Amadalvarez (talk) 22:21, 3 December 2022 (UTC)[reply]
Ah, sorry for the confusion, I am adding them to the candidates themselves, i.e. Mark Rutte (Q57792). So that's not necessarily related to the infobox (I think?). Dajasj (talk) 22:54, 3 December 2022 (UTC)[reply]
OK. Remember that in these items we have election results only. Not the finally composition of chambers, nor position held (P39) of elected, etc.
When I wrote the ontology I analized how we may use the results info to create (via bot or similar) the P39 or the list of parliament members, it is, the names of the number of seats in assembly (P1410). But it requires to enter the info of all persons for all lists (if we do before election day) or the final list names (in post-election day), in some place (may be participant (P710)), but it is against best practices of data model: Don't have long lists (in the Netherland, 150; in Spain, 350) and build lists via inverse access, it is SPARQL.
So, the content of election item are just "election results".
Regarding using Infobox, you may propose in your chat using the example pages.
Cawiki community were not different to other editors, but.. when they saw how much generated a simple call without parameters, it was easier to change they point of view.
In cawiki, 85% of whole articles have infobox feed from WD. Amadalvarez (talk) 23:29, 3 December 2022 (UTC)[reply]
Okay, but so for candidacy in election (P3602) using represents (P1268) instead of member of political party (P102) isnt against any rules, right? Dajasj (talk) 21:31, 5 December 2022 (UTC)[reply]
@Dajasj, Not to me.
P3602 is a property for person item and, in my opinion, to hold info of elections differents than generate a P39. It's, when the candidate did not get position, the P3602 allow us to know that was a candidate; however, some other people thinks that P3602 must be a summary of all his/her P726.
In person item, I believe that P1268 are equivalent to P102. The difference is that P102 is a main property for a person, when P1268 is not a personal characteristic, but in relation to something; so, should be a qualifier of P3602 or P39.
Does it respond to your doubt ?. Amadalvarez (talk) 22:38, 5 December 2022 (UTC)[reply]
Yes thanks, that solves it! And I guess I am one of those people adding all those P3602, because a lot of candidates don't get elected (or decline taking office) ;) Dajasj (talk) 07:37, 6 December 2022 (UTC)[reply]
Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata_talk:WikiProject_elections&oldid=2070311299"