User talk:Owain/archive2

Unexplained reverts

Hi Owain, you seem to have mass reverted edits without any explanation. Somerset (Q67461071) is instance of (P31) historic county of England (Q1138494) which is a subclass of (P279) historic county of the United Kingdom (Q67376938). Adding historic county of the United Kingdom (Q67376938) directly to Somerset (Q67461071) is redundant and not how Wikidata works. Also it's considered bad behavior to (mass) revert without any clear explanation. So please explain yourself. Multichill (talk) 18:12, 15 May 2020 (UTC)

Sorry, that was in error. Owain (talk)

Sports venues

Excuse me Owain, but what's the point in adding the historical country in the items of the sports venues of the United Kingdom? -- Blackcat (talk) 08:36, 6 August 2020 (UTC)

Not 'historical', but 'historic'. Any place, building or structure falls within three separate geographies (county, lieutenancy, local government) as shown by the ONS Index of Place Names: https://geoportal.statistics.gov.uk/datasets/index-of-place-names-in-great-britain-july-2016/data?geometry=-4.269%2C55.823%2C-4.236%2C55.828 The bot simply adds the county to any item that has a local government area defined. Owain (talk) 08:43, 6 August 2020 (UTC)

But is it useful for searching purpose or similar? -- Blackcat (talk) 17:59, 6 August 2020 (UTC)

Of course, to find all stadia in Renfrewshire:

SELECT DISTINCT ?item ?itemLabel ?coord WHERE {
 ?item wdt:P7959 wd:Q67545499;
   wdt:P625 ?coord;
   wdt:P31/(wdt:P279* wd:Q483110.
 SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}

There are countless ways to search by county, e.g. Listed buildings in Lanarkshire:

SELECT DISTINCT ?item ?itemLabel ?coord WHERE {
 ?item wdt:P7959 wd:Q530296;
   wdt:P625 ?coord;
   wdt:P1435 ?heritage.
 SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}

And so on... Owain (talk) 18:22, 6 August 2020 (UTC)

We sent you an e-mail

Hello Owain/archive2,

Really sorry for the inconvenience. This is a gentle note to request that you check your email. We sent you a message titled "The Community Insights survey is coming!". If you have questions, email surveys@wikimedia.org.

You can see my explanation here.

MediaWiki message delivery (talk) 18:46, 25 September 2020 (UTC)

Scottish parish coordinates

Thanks for adding coordinates for Scottish parishes.

It might be nice to add a criterion used (P1013) determination method (P459) qualifier to indicate what the coordinate point represents, because various choices might have been considered -- eg centroid (Q511093), average of cardinal extremal points, focal point of largest population centre, etc -- so not a bad idea to say which. Jheald (talk) 15:15, 19 November 2020 (UTC)

The batch name did mention that it was centroids from the NRS shapefile, but that seems to have been missed off. I will create a new batch to address the reference when I have tidied them up a bit. Owain (talk) 15:18, 19 November 2020 (UTC)

So now I see you are removing all of the existing coordinates, added by people like User:James nayler is real and others. Please be careful about things like this, and discuss substantial removals of content, if possible with the people who added it. In this case, it looks like many of the coordinates you are removing were coordinates for the parish church associated with the parish -- which may be far from the worst way of locating the parish; and also information valuable to have in its own right. Yes, there are other ways that information could be represented, eg parish church (P8289) linking parish to church, and then making sure the church has coords. But P8289 isn't well populated yet, and doing so accurately is not without its traps (eg different churches at different periods of time; and only some civil parishes still corresponding to ecclesiastical parishes), so please be cautious before you remove information like this. See also en:Chesterton's fence. Jheald (talk) 18:46, 24 November 2020 (UTC)

I understand that, but an item can only have a single P625 value, and the centroid of each feature in the NRS shapefile is the only data that can be accurately referenced. Once each item has a single P625 I will batch add the reference. The P625 of the parish churches is obviously vital data too, and I will certainly assist in adding it to each relevant item where missing. Owain (talk) 18:51, 24 November 2020 (UTC)

No. Items can have many values for any particular property, distinguished by different sources, different determination methods, different roles, etc. Jheald (talk) 19:53, 24 November 2020 (UTC)

Thanks for tagging me in this thread. It's good to know that more data is being added to Scottish parishes. I'd argue that the irregularity of Scottish parish shapes means that centroid data from NRS shape files is insufficient to capture the idea of a parish and can result in some strange anomalies (see Hamilton). In my minor edits, I had attempted to find an approximate centre point and then use the other coords (for N, S, E, W bounds) to hedge the limits of the parish. I appreciate this is not scientific, but it does capture approximate centre points of parishes despite potential changes to their boundary locations over time.James nayler is real (talk) 09:25, 26 November 2020 (UTC)

>>>>>>>>>Hi Owain. I've been thinking about your decision to use centroid data in adding parish coordinates. I came across an example that I believe shows how this approach is suboptimal. The entry for Cramond (https://www.wikidata.org/wiki/Q74762146) now includes coordinate locations that show the parish somewhere in the middle of the Firth of Forth.

This is considerably devaluing this data for other users (I've had to manually correct the coords in my own download of this data). I don't want to make any changes to the Q file without coming to some agreement with you and Jheald so we can move on with our projects. Is there some compromise we can devise instead of using centroid data, please?

Yes, I have experienced the centroid-out-of-polygon problem myself with very irregularly shaped polygons. The problem with Cramond is most of the ancient parish was swallowed up by the city parish of Edinburgh, so that the only bit of the civil parish left is Inchmickery and Cow & Calves (the latter of which is not even in the NRS polygons). The current centroid is actually right in the centre of Inchmickery, which is correct if we exclude Cow & Calves, but it does not show in the Wikimedia map. Are there any others which are out-of-polygon? Owain (talk) 11:53, 3 December 2020 (UTC)

>>>>Is there any way we could add the centroid coordinates under a different property? P5140 would allow you to show how you devised the centre point.

While I understand that the centroids can be referenced, the name of the parish and the settlement in which it sits seems to be the most logical choice for a centre point coord, and unlikely to produce these anomalies.James nayler is real (talk) 14:28, 3 December 2020 (UTC)

That sounds like a good enough compromise, as long as there is a named location. There is sometimes a Kirkton of X or the like, where the parish name does not relate to a village. Perhaps parishes like Assynt where there is no village, and no Churchtown of Assynt would need to retain the centroid? Owain (talk) 14:36, 3 December 2020 (UTC)

Happy to retain the centroid where the location isn't referenced in the parish name. I think P5140 will allow you to continue to upload the geographic centre of the parish to each item. You're quite right in saying that these are important calculations. I think P5140 would be a more suitable home for these properties.If this is all good, would you be able to revert the coords back to their originals when you add the P5140 centroid coords, please? James nayler is real (talk) 15:07, 3 December 2020 (UTC)

Thornton Cemetery

You do know there are other THornton Cemeteries, I take it? So why lose the important disambiguator? I'm going to revert that. Rodhullandemu (talk) 15:05, 22 March 2021 (UTC)

The location is defined by properties, not the title of the entry, which should be just the proper name without disambiguation. Owain (talk) 15:07, 22 March 2021 (UTC)

That's OK when we have fully knowledgeable AI that understands numbers. If that's a rule, it's a stupid one and should not be followed slavishly. Meanwhile, these things have to be maintained by us poorly equipped humans, who don't want to search through a list of identically-named alternatives to find the right one. Sorry, I've worked in the real world all my life, and I don't intend to become unreal for the remaining couple of years of it. Rodhullandemu (talk) 15:13, 22 March 2021 (UTC)