Wikidata talk:WikiProject British Politicians/Archive 2

Suspension of Ian Paisley, Jr

@Andrew Gray: I see you've already created a new P39 for Ian Paisley Jr (Q575881) being suspended from his party. What's your thinking on recording his suspension from Parliament itself? We do have an item for the concept — Suspension from the UK parliament (Q7649223), via its enwiki page — though currently nothing is using it. --Oravrattas (talk) 16:35, 28 July 2018 (UTC)

@Oravrattas: I plan to do this but there's a bit of a weird snag - Paisley's suspension is in the future. As I understand it, he will be suspended when Parliament resumes in September - he's not suspended now, because otherwise they'd be in the situation of suspending him when Parliament wasn't in session anyway, making it a bit of a limited punishment! I've held off making these edits for now because I don't want to mess around too much with future dates.

Having said that, it seems we don't actually model any of the people at w:Suspension from the UK parliament - I had thought we had for Jonathan Sayeed (Q6274334) at the very least. I will add this to the to-do list - end one term on day of suspension with appropriate qualifier, start a new one X days later, as though they'd very briefly resigned and been re-elected. Probably be a way down the priority list, though! Andrew Gray (talk) 21:34, 28 July 2018 (UTC)

Ah, I hadn't quite realised that the suspension didn't take effect straightaway. So, yes, seems sensible to wait until it's actually in force. --Oravrattas (talk) 08:24, 29 July 2018 (UTC)

That said, I don't really know what we should do come September. End the current term and leave it there - but that risks forgetting to update it on his return. We could add his return dated in the future, but that comes back to the same problem of "things that haven't happened yet". One to think about! Andrew Gray (talk) 10:01, 29 July 2018 (UTC)

Constituencies "said to be the same as" earlier versions

The suggestion to use said to be the same as (P460) to tie together historic versions of constituencies seems a little odd to me. My understanding of that property is that it's used for articles that we'd like to merge, as they're believed to be conceptually the same thing, except there's either two different Wikipedia articles in one language, so we practically can't merge, or someone actually disputes that the two things really are exactly the same. Wouldn't this be closer to different from (P1889) (these might look at first glance like they're the same thing, but they're actually conceptually separate)? --Oravrattas (talk) 11:40, 14 September 2018 (UTC)

Comparisons against Parliament API

@Andrew Gray: I set up a proof-of-concept prompt for the list of the current MPs in Wikidata vs the Parliament API at User:Oravrattas/sandbox/prompts/UK57. As you can see the vast majority of the differences are simply down to different versions of names for people, parties, and constituencies, all of which disguise the actual differences we'd want to look at — primarily around which MPs are now/no longer Independents. I can think of a few possible approaches here:

1. Either the SPARQL or SQL could be full of special cases to remap, e.g. "Conservative Party" to "Conservative" or vice versa. This would get unwieldy very quickly, but might have some quick wins whilst look at better solutions.
2. The party data from the Parliament API includes IDs. If these are stable and consistent over time (which I'm not sure about, but we could investigate/ask), we could potentially create a new Property for these and then match on those instead.
3. We could add some sort of "name at source" value to items qualified as saying this is how the Parliament API refers to them. Then the SPARQL to compare against this source could explicitly prefer any names specified that way (with a fallback to the label if none is provided).

The third is probably my preference out of these, but we'd also need to decide which way around to do this. For example, we could use described at URL (P973) with a subject named as (P1810) qualifier, as used in the South African Parliament data: Q16744297#P973; or we could use name (P2561) (or one of the more specific variations), with either a reference URL, or some more specific qualifier, to 'tie' this version of the name to the Parliament API.

I think the value of having daily comparisons of where our data differs from the Parliament API makes it worth finding a good solution to this. Thoughts? --Oravrattas (talk) 11:13, 17 January 2019 (UTC)

@Oravrattas: This is great and I love it. Definitely agree it's valuable to have.

I think the best solution is #2 - it looks like you're currently matching on the numeric part of parliament.uk biography pages (BEING DELETED) (P1996), but UK Parliament ID (P6213) is broader in scope and can handle parties & constituencies. It's already in place (I think) for all current MPs so I could extract and upload it for constituencies very straightforwardly (eg Hackney N). (The formatter URL is currently on the "pretty" biography view for people, but any id will resolve at id.parliament.uk so perhaps we could switch it). Andrew Gray (talk) 12:38, 17 January 2019 (UTC)

@Andrew Gray: Oh, excellent. I didn't realise we had a suitable property that would work across all everything. Comparing IDs is always a lot better than comparing strings! (There's a slight awkwardness in that the resulting table will be fairly impenetrable if it only contains numbers, but there's probably something we could do with custom templates </handwave>. I'll play around/give it some more thought.) Do you happen to already know how to get a list of current members out of the system with IDs under that scheme? --Oravrattas (talk) 14:12, 17 January 2019 (UTC)

OK, looks like my best bet might be https://api.parliament.uk/query/house_current_members.json?house_id=1AFu55Hs — I'll have a play with that and see where I end up. --Oravrattas (talk) 14:23, 17 January 2019 (UTC)

User:Oravrattas/sandbox/prompts/UK57 vs API is the new version, using that approach (based on https://morph.io/tmtmtmtm/uk-parliament-api-members — I'm not entirely sure the party membership bit is guaranteed to work for everyone in future, but it seems to do the right thing for now). So next we need to have the relevant IDs added to everything. I'll do the parties by hand, and if you can add the constituencies, then we can remove the text fields, and, in theory, it should alert us only to significant differences. Then I'll work out how to make the output more meaningful --Oravrattas (talk) 15:08, 17 January 2019 (UTC)

Hmmm, Independents are a slightly awkward case now. I'm not so sure about setting

⟨ independent politician (Q327591)    ⟩ UK Parliament ID (P6213) ⟨ UeAtkxYY ⟩

Perhaps we need a new item specifically for being Independent in the UK? --Oravrattas (talk) 15:17, 17 January 2019 (UTC)

@Oravrattas: So, constituencies. they seem to have both "HouseSeat" and "ConstituencyGroup" IDs. For Abbott, those are 0BhROnYP & 5bX5Se0u, neither of which seems to match to HSoMS1VX, the one I thought represented the seat.

5bX5Se0u (ConstituencyGroup) seems to be the one most closely related to what we'd think of as a constituency item, and I *think* what's going on here is that it represents Hackney N since 2010, while HSoMS1VX is 1997-2010 (there was a small boundary change and it picked up some new wards). We would be unlikely to model the relatively small difference between these two in Wikidata, at least not in the near future - would your SPARQL be able to report a match if there were two values? (We could potentially add start/end dates to the IDs so that it could successfully pick out the current one, of course)

On "independent", maybe it would be best to code this one as an exception - it feels a bit odd to declare that being independent in the UK is different from elsewhere. Andrew Gray (talk) 21:23, 17 January 2019 (UTC)

@Andrew Gray: agreed on 5bX5Se0u. I would need some way of choosing the correct P6213 if there were more than one on an item, but having start/end dates would certainly be enough. I'm not sure what your process for adding these looks like, but even if all we had for now were the current constituencies with their current ConstituencyGroup IDs, I think that would be very useful, and it would be simple to update the prompt later if we need to refine it to cope with historic IDs.

I do agree that it initially seems slightly odd to have a separate item for UK Independents. But here we have an external ID property with a value for just that! Special casing it in our code means that anyone else who wants to work with this data would also need to know to special case it in their code as well. Essentially the existence of this ID means I think it's fine to treat this as a quasi Parliamentary Group of the House of Commons, and from that angle it seems perfectly fine for it to have its own UK-specific item: indeed it would almost be odder for it not to. --Oravrattas (talk) 10:46, 19 January 2019 (UTC)

@Andrew Gray: I realised this afternoon that the comparison itself can actually give us the perfect CSV file of constituency items to the parliament IDs, so I converted that to QuickStatements format and uploaded them all.User:Oravrattas/sandbox/prompts/UK57 vs API now only shows actual differences. Most of these are the independents, but there are a few other party issues there. The page is a little bit impenetrable, as it only shows the Parliament IDs for everything, but hopefully it's easy enough to see the problems and work out what to do. If we wanted something a little more user-friendly we could conceivably also include a fake column in the CSV for the Wikidata IDs of the party and/or constituency, and supply custom templates that only display them where there's actually a difference. It would be a little bit fiddly, but shouldn't be too difficult. See what you make of this version, and if I get some free time over the next week or so, I'll see about making it a bit nicer. --Oravrattas (talk) 15:50, 20 January 2019 (UTC)

Independent Labour/Conservative

@Andrew Gray: The documentation on "Removal of the whip or resignation from the party" currently states: "If they leave the party to sit as a "named independent", it may potentially be worth using a more specific qualifier than parliamentary group (P4100):independent politician (Q327591), but be careful that this is not actually a party - for example, Clare Short (Q333550) sat as "Independent Labour", which was not the same as Independent Labour Party (Q1507913). For the moment, all independents are simply listed as parliamentary group (P4100):independent politician (Q327591)."

This came up again this morning for me as when comparing some more of our data against the Parliament API, I traced a few more of the differences back to this category of Independents. The Parliament API does have a separate item for "Independent Labour" (UK Parliament ID (P6213)=65F7fNUX), applied to Clare Short (Q333550) and about 15 other people. Similarly, for Ann Winterton (Q337529) and about 25 other people, Parliament has "Independent Conservative" (0uBsH1dr). Should we go ahead with creating items for each of these, and adjust all the parliamentary group (P4100) qualifiers accordingly? --Oravrattas (talk) 09:16, 5 February 2019 (UTC)

@Oravrattas: Good question. I think one of my concerns here is that linking them together as "independent Labour" gives the impression that that's a coherent party grouping - though tbh you could also say the same for putting them all as "independent". Does it have anyone explicitly listed as purely "independent"? Andrew Gray (talk) 20:38, 5 February 2019 (UTC)

@Andrew Gray: Yes, there are quite a few that are plain "independent" — e.g. Sylvia Hermon (Q676383) and Fiona Onasanya (Q30163918). I think if we were to make "independent Labour" etc an instance of independent politician (Q327591), rather than of political party (Q7278) or parliamentary group (Q848197) or the like, it would be relatively easy to still filter those differently if required, and, as you say, pretty much any issue with doing it that way would also apply to simply having them grouped with independent politician (Q327591) anyway. Plus, of course, this way is what the official source actually says. --Oravrattas (talk) 20:48, 5 February 2019 (UTC)

@Oravrattas: I'm a little cautious here because I'm not quite sure why those categories are used in each case - is it a self-identification thing? - but if it's the ones we're getting from Parliament, then let's go for it for now. As long as they're defined as subsidiaries of "independent politician", we should be fine. Andrew Gray (talk) 21:31, 5 February 2019 (UTC)

Modelling the start cause for a term

At the moment, there is no way I can see which people commonly use for representing the reason a given P39 has started.

Is there precedent anyone has come across for using has cause (P828) as a qualifier on P39s with a value such as joining party (Q52084147), or would this have to be inferred from other data where available?

--jacksonj04 (talk) 09:38, 24 April 2019 (UTC)

• Well, there is "elected in", which does the job 95% of the time :-). But I agree - for parties you can potentially infer it from the previous P39, but for things like replacement list AMs/MSPs, that doesn't really work. has cause (P828) seens the logical counterpart to end cause (P1534), though oddly the properties aren't clearly marked as such.

I've been thinking about how to handle this for the Lords, as well. No elections (usually) so we'd want to model "start cause" with values like inheritance, new creation, writ of acceleration, etc. I assume other parliaments with members who are appointed by different means have similar issues but I don't know if we model any yet. Andrew Gray (talk) 12:53, 24 April 2019 (UTC)

• • For MEPs, one might want to think how to model start cause = death or resignation of someone higher up the election list. Jheald (talk) 13:43, 24 April 2019 (UTC)

End cause for non-Parliament P39s

I'm currently looking at data for member of the London Assembly (Q56573014), and wanted to canvas opinion on the best way to describe the end cause for those terms.

As an example, Andrew Dismore (Q338076) has served two consecutive terms in the London Assembly with no party or constituency changes. We'd usually model these as two distinct P39 statements - in the case of memberships of Parliament the end cause (P1534) is set to dissolution of parliament (Q741182), and the description of this implies it would apply to any predefined end of a legislative term, but it doesn't feel quite right to use.

Does anybody have any views on the best end cause for these non-Parliamentary P39s?

--jacksonj04 (talk) 09:43, 24 April 2019 (UTC)

New item for "end of legislative term" ? Jheald (talk) 10:30, 24 April 2019 (UTC)

Agreed, this seems best. As noted, it's basically the same concept but no harm in having a more specific one for when there isn't a dissolution per se. Andrew Gray (talk) 12:49, 24 April 2019 (UTC)

Seems solid to me - created as end of legislative term (Q63323711), in use on Andrew Dismore (Q338076) (and I'll use on the rest of the London Assembly members), and dissolution of parliament (Q741182) has had its subclass changed to the new item. --jacksonj04 (talk) 09:15, 25 April 2019 (UTC)

End cause for non-Parliament P39s

I'm currently looking at data for member of the London Assembly (Q56573014), and wanted to canvas opinion on the best way to describe the end cause for those terms.

As an example, Andrew Dismore (Q338076) has served two consecutive terms in the London Assembly with no party or constituency changes. We'd usually model these as two distinct P39 statements - in the case of memberships of Parliament the end cause (P1534) is set to dissolution of parliament (Q741182), and the description of this implies it would apply to any predefined end of a legislative term, but it doesn't feel quite right to use.

Does anybody have any views on the best end cause for these non-Parliamentary P39s?

--jacksonj04 (talk) 09:43, 24 April 2019 (UTC)

New item for "end of legislative term" ? Jheald (talk) 10:30, 24 April 2019 (UTC)

Agreed, this seems best. As noted, it's basically the same concept but no harm in having a more specific one for when there isn't a dissolution per se. Andrew Gray (talk) 12:49, 24 April 2019 (UTC)

Seems solid to me - created as end of legislative term (Q63323711), in use on Andrew Dismore (Q338076) (and I'll use on the rest of the London Assembly members), and dissolution of parliament (Q741182) has had its subclass changed to the new item. --jacksonj04 (talk) 09:15, 25 April 2019 (UTC)

MPs as Special Advisers (and other jobs I suppose)

Someone raised this the other day and I thought it might be a good area for enrichment. I've added Ed Miliband's SPAD roles using the "of" qualifier to denote who they were advising.

Any ideas for how this could be done better, or does this seem ok for now? Battleofalma (talk) 10:51, 1 July 2019 (UTC)

@Battleofalma: sorry I missed this comment! I think this looks good - my only qualm is whether SPADs are technically appointed to advise a minister (eg Brown) or a department (eg the Treasury). Either way I think we could definitely make good use of this. Andrew Gray (talk) 09:30, 18 August 2019 (UTC)