Talk:List of built-up areas in Wales by population

Wrexham

We are not treating any of the wider urban areas as a single unit (cf. Cardiff Urban Area and Swansea Urban Area) so there's no case to treat Wrexham any differently. Maybe this uncertainty is partly my fault for moving the page to "List of urban areas in Wales by population"; I will move it again to "List of settlements in Wales by population" if that's helpful. There is a lot of confusion between the ONS jargon of settlements, urban areas and urban subdivisions; I *think* that settlement is actually an exclusively Scottish and Northern Irish term, with the less intuitive urban subdivision applying in England and Wales, but given the use of List of largest United Kingdom settlements by population I think it should be OK here. Pondle (talk) 17:10, 28 May 2009 (UTC)

Attack of the Zombies

I don't think the data-set selected in its current form is particularly useful for comparing the populations of settlements in Wales. If you look in National Statistics' glossary [1], you'll find the definition of "Urban subdivisions":

Urban subdivisions

Major urban areas and others with more than one central focus are divided where possible to produce figures about localities within them. Previously separate urban areas, where urban land has merged, are also recognised by subdivisions where possible. Subdivisions often follow the boundaries of local authorities existing before reorganisation in 1974, or the boundaries of current authorities within urban areas.

Unfortunately in certain parts of Wales the local government boundaries pre-1974 did not reflect the settlement pattern having changed from rural to urban as a result of the Industrial Revolution. This results in several anomalies in the data-set, for instance:

4) Rhondda -- This is the old Rhondda MB, which consists of two valleys with a number of settlements dating from the second half of the 19th Century (before the 1850s, the population was less than 1000). There is of course no such place as "Rhondda town centre". It was not until 1974 that the co-extensive parish was divided into 16 communities.

19) Shotton/Hawarden -- This is the old Hawarden RD. This is possibly the worst of the lot. Not only are the two places named very distinct from each other, but also Shotton abuts (32) Connah's Quay directly across the Wepre Brook, with a continuous main street! This case doesn't get much better at community level: Hawarden Community hides four settlements: Ewloe (the largest), Hawarden, Aston (directly adjoining Shotton), and Mancot. Short of resorting to Super Output Areas, all statistics relating to this area are to some extent misleading.

22) Mountain Ash / Abercynon -- This is the old Mountain Ash UD. As the name suggests, the urban subdivision resembles a dumb-bell with a very thin bar.

24) Pontardawe / Clydach -- Pontardawe RD. Once more, it's the stereotypical dumb-bell.

29) Brymbo / Gwersyllt -- Wrexham RD. This one doesn't look too bad until you take into account that the topography (the Moss Valley is between the two) makes both of these much more closely linked to Wrecsam than to each other.

31) Gelligaer -- Although there's a history involving Gelligaer UD, the settlement in question is better known as Ystrad Mynach, although the area also includes Hengoed, Maesycymer, and Y Gelli-gaer.

I could go on, but those are the obvious problem entries above 15,000. Roughly one settlement in five in this sample is in some way misleading. The bigger problem is that there is no dataset that shows what we want. Urban areas are even more prone to grasping places with tendrils (Neath (8), for instance, disappears into Swansea (2)). Communities are extremely variable and inevitably include rural areas, not to mention the fact that there are no communities called Cardiff, Swansea, Newport, and Wrecsam.

I really don't know what to suggest. There is some use for this article, so I don't want to propose deletion. It would clearly be original research to try to remedy the massive flaws in the dataset. Would somehow noting the flaws be something that could gain consensus? 𝐨𝐱𝐲𝐩𝐡𝐞𝐧𝐵𝑈𝑇𝐴𝑍𝑂𝑁𝐸 ⓊⓉ 19:50, 17 September 2009 (UTC)

I'm a little confused by your post because the ONS-defined 'urban areas' aren't based upon old council boundaries, they are derived by (this is an abridged version) identifying output areas with permanent structures extending for 20 hectares or more, and with more than 1,500 residents. Separate areas of urban land are linked if less than 200 metres apart. There might be some approximation to old urban districts but I don't believe that it's intentional. More information here and here Pondle (talk) 20:48, 17 September 2009 (UTC)

I can see what you're being confused by. Firstly, this article is using urban subdivisions, not urban areas (except where an urban area is unsubdivided). These are defined as per the document I quoted in my original message. Secondly, the PR PDF you linked to obfuscates this definition (it observes that modern boundaries are not necessarily used in subdivisions, but does not admit their true nature).

I do believe there are good reasons for this article to continue not to use National Statistics' urban area figures: the 200-metre rule in particular does not play nicely with the form of urbanism in former mining areas, although it is at least consistent. The urban subdivisions are a messy compromise between a geo-historical definition and a statistical one and therefore have messy flaws. 𝐨𝐱𝐲𝐩𝐡𝐞𝐧𝐵𝑈𝑇𝐴𝑍𝑂𝑁𝐸 ⓊⓉ 22:19, 17 September 2009 (UTC)

Ah sorry, I see what you mean re: subdivisions. This article is treating the urban subdivisions as separate entities as per the precedent at List of largest United Kingdom settlements by population. Although you may have a point about flaws in the ONS methodology, I don't see a case for change on the basis of opinion; the criteria for inclusion is WP:V and the ONS is a reliable source. If however the methodology for deriving ONS urban stats has itself been criticised in other reliable sources, then we could definitely add that as context. Pondle (talk) 22:38, 17 September 2009 (UTC)

Urban subdivisions are probably the least understood (and so most contentious on Wikipedia) ONS census dataset, and especially their link to pre-1974 local government units. There have been many claims on Wikipedia that they're simply recycling pre-1974 units, which is false. What is true is that within conurbations, the boundaries between urban subdivisions usually follow the pre-1974 boundaries, simply because post-1974 many towns lost their individual identity as far as local government was concerned. Outside conurbations, the ONS use the 200m land use rule to identify individual settlements - and where two settlements were part of the same pre-1974 local authority, and were linked at the time of the 1981 census, then you get the strange Mountain Ash / Abercynon type naming convention, as neither has primacy over the other and it's impossible to separate them as there's no pre-existing boundary to use. Where settlements have expanded into former rural areas from the previous census (or in the 1981 census into former Rural Districts), then those areas are considered to be part of the settlement.

As it happens, Mountain Ash/Abercynon isn't particularly dumb-bell shaped, but is more of a linear form. Fingerpuppet (talk) 21:59, 18 September 2009 (UTC)

Nefyn & Rhondda

Nefyn article has this: Nefyn is a small town and community on the north west coast of the Llŷn Peninsula in Gwynedd, Wales, with a population of 2,619.
Still, it is missing from the list. Is there a reason it doesn't belong on the list, or does it belong to some other entry? 85.217.46.52 (talk) 21:49, 1 September 2012 (UTC)

And, List of largest United Kingdom settlements by population has "Rhondda" at position #140 with population of 59,602. It is missing from this list. Reason? 85.217.46.52 (talk) 21:54, 1 September 2012 (UTC)

You're absolutely right about the Rhondda - but it means updating the whole rank column, I had a look for some tips but frankly life is too short! As for Nefyn, I can't find it in Table KS01, which is the source for the page, although that hasn't stopped people from adding numerous other villages to this list. Pondle (talk) 20:59, 4 September 2012 (UTC)

Parts of larger urban areas which are mostly in England are missing from this list also notably Saltney and Broughton, Flintshire. Also Nefyn's population of 2,619 is the population of the parish not the urban area which this list uses. This article's source doesn't mention Nefyn at all. They also have maps of these on the ONS website with the urban areas on them and Nefyn isnt on their either (I checked in case it was listed under a different name). Eopsid (talk) 18:44, 10 September 2012 (UTC)

Hmm, how about Rhondda then? Still not in the list. However, some cities/towns from its area (Rhondda Cynon Taf) are on the list. Don't know if it is the same because there are additional words, Cynon & Taf. 85.217.44.150 (talk) 02:06, 30 January 2013 (UTC)

Where is #40?

Hi there! Your list seems to be missing a position in that the table goes straight from:

39 Carmarthen 14,648 Carmarthenshire

to:

41 Abergavenny 14,055 Monmouthshire.

Is there actually a missing town or is it just that someone can't count? :-)

Or perhaps I'm blindly overlooking something else which makes this gap okay (it's been known to happen!) All the best --ThunderingTyphoons! (talk) 10:31, 31 May 2013 (UTC)

If there something which makes it okay, it certainly is not mentioned. That would be least to expect in that situation. I bet it is just an error. 85.217.42.90 (talk) 13:13, 6 August 2013 (UTC)

Upon further studying, I see the above-mentioned Rhondda is missing from the list. Apparently, it was removed in this edit: [[2]], a village "Ysbyty Ifan" was added, and Wrexham moved to Rhondda's place at #4. 85.217.42.90 (talk) 13:35, 6 August 2013 (UTC)