A puzzle for Xmas .... which WHS are likely to be in such a list? What size of population might be expected? The information is surprisingly lacking - one might have thought that it would have been an essential figure to be included in Nomination Files and Reactive Monitoring Mission reports - but it is not - so we are faced with a problem of trying to find or estimate it.

Candidates are likely to be cities whose inscribed area is not all "public buildings" or ruins etc but includes a significant currently inhabited area - though they could also include large Cultural Landscapes which include a number of complete towns. Identifying the population of either requires knowing or discovering the inscribed area in both size and nature (ie how much is "built up"/populated). Titles can be misleading, e.g sometimes giving an impression that they might include a far larger area than they do - The core area of "Kathmandu Valley" for instance only includes a mere 167.37 ha of the "Valley" - much of which is going to consist of temples etc.

The "inscribed area" is (usually!!) easily available so could provide a useful proxy for population to help filter out the likely candidates. What might one expect in terms of population densities per hectare? (1sq Km = 100 ha). A few indications from the Web -
"A population density considered "heavy" is generally over 80 people per ha"
"Dense urabn areas in major cities like Manhattan might reach densities of 180 per ha while suburban areas might have densities closer to 20 people per ha"
"Neighbourhoods with residential densities of well over 80 persons per ha (above the London average is 68.6 pers per ha) can be found in typical outer suburban areas such as Croyden, Ilford..."
"A heavy population density for an Indian city would be considered anything above 20000 people per Sq Km (= 200 per Ha). Cities like Mumbai and Kolkata often reach this level or even surpass it"

The first WHS where I have been able to establish a published and reasonably up to date figure for the population of the Inscribed area is Brazilia. The inscribed area is 11,268.92 ha (UNESCO Web site). I quote from the 2012 "Joint WHC / ICOMOS Reactive Monitoring Mission Report, Brasilia" - "The area inscribed on the World Heritage List corresponds to the Plano Piloto itself."......"If the population of Brasilia was 198,000 in 2000 (since Cruzeiro and Candangolandia are located within the inscribed property), the 64,000 inhabitants of Cruzeiro and the 15,000 inhabitants of Candangolandia should be added to this figure, supposing a total below 300,000 inhabitants), the Plano Piloto has hardly grown ten years later (always including the nuclei situated within the inscribed property) up to 304,000 inhabitants.".
304000 people spread across 11269ha is approx 27 people per ha. not reaching the "Heavy population density" levels but not seeming unreasonable for a modern city with very large areas of space given over to Federal Buildings, open areas etc but also one where much of the population lives in modern apartments. Although having a relatively light density Brazilia gains in population because of the rather large inscribed area when compared with that of many "Historic Centres" which might have a greater population density but in a much smaller inscribed area
So - I propose Brasilia as a benchmark for the "WHS with the largest population" at 304000 in 2012!!

Another likely "most populated" is Rome. But how to establish the population of the inscribed area? ( Map here)
Rome is divided into 15 "Municipi" - See map of them here.
The inscribed area of Rome is significantly (but not entirely) contiguous with the boundaries of Municipio I - see its map here
I think it can be agreed from the respective map shapes that ONLY Rioni 22 (Prati) has been excluded from the WHS (it only got included in Municipio I during a reorganisation in 2013) whilst the very small San Paolo Fuori le Mura needs to be added.
Wiki gives the current area of Municipio I as 1991 ha. I am unable to identify the area of Prati but the "UNESCO stated" inscribed area of Rome as 1430.8 ha plus Holy See territories of 33.57 and 5.33 is compatible.
Wiki gives the population of Municipio I in 2019 as 167,330. And that of Prati as 20538 = a population of the inscribed area of 146792 (At a density of c102.6 - reasonable for a packed city centre with a lot of historic buildings?)

So we have (approximately but good enough?)
Brasilia 304000 (2012)
Rome 147000 (2019)

Marrakesh 167000 (2004) Assume that, although the entire city of Marrakesh has grown in population since then, this has not impacted the Medina which was historically "packed".
The site consists of 3 elements - the Medina, the Agdal Gardens extending from it and the Menara Gardens separately to the West. Assume the gardens have no significant population (justified by maps and satellite photos)
This report describing the "Commune de Marrakesh" contains size and population details of the various parts of the city. In Its figures it states "La ville de Marrakech est organisée en cinq arrondissements, présentant chacun des caractéristiques spécifiques du point de vue de sa population et de l'organisation de l'espace. Tableau 2 : Présentation synthétique des arrondissements......"
The Arrondissment of Medina covers 570ha and had a population of 167,223 with a density of 293 per ha. (elsewhere the significant density of the Medina is commented upon and is compatible with its "inscribed nature"?)

UNESCO states that the WHS covers 1107ha. Wiki states that the Agdal Gardens cover 340 ha and the Menara Gardens are around 90 to 100 ha plus the Medina itself at 570 = 1010ha. so only a 100ha unaccounted for

Maybe the exceptionally beautiful city of Ahmedabad, the population of the nominated property is 370,000 mentioned in the ICOMOS evaluation report.

Medina of Tunis 110000
"With an area of 270 hectares (670 acres) (plus 29 hectares (72 acres) for the district of the kasbah) and nearly 110,000 inhabitants, the Medina has one-tenth of the population of Tunis and a sixth of the urbanized area of the agglomeration." (Wiki) https://en.wikipedia.org/wiki/Medina_of_Tunis
UNESCO has the core area at 296.41ha - so, a population density of 341.

With Durian's Ahmedabad 370000 across 535.7ha (an amazing 653 pop density) we now have 5 WHS for the Connection. I believe there are more (though whether we can "prove" them is another matter) so suggest that we place a lower bar at 100000 inhabitants in the core zone?

Would we want to exclude serial sites covering multiple cities? Was thinking of the Great Spas of Europe. Bath has a fair number already.

Bath actually is quite unusual in including so many residential areas (many of which are of no particular historic or cultural note whatsoever!) so, in one respect, it illustrates the point. They were determined to use exactly the same area as that of their own separate WHS even though the justifications were different. It alone with 2570ha represents over a third of the inscribed area of 7018ha of the Spa WHS across 11 towns with only Karlovy Vary (1123ha) also having a significant area and probably population. Bath's 94000 population would certainly seem likely to push the Spa WHS over 100k in total. I am not sure that many other "serial sites covering multiple cities" would also do so. Such WHS do create a problem in that they multiply the problem of identifying the residential populations!

Further to this - it is noticeable that most of the European capitals which have been inscribed have done a "good job" at restricting the core to a genuine small "historic" area which is also largely full of monuments, government buildings etc and therefore doesn't have a large resident population
I have had a look at a few
Vienna - only 23000 live in the core area of 371ha - a density of 62
Prague - Leaving out the extension to include Pruhonice Park (which won't have many inhabitants) around 44200 live in the originally inscribed area of 894ha - a density of 49.5
Amsterdam - the Canal Ring area only covers 198.2ha but one might have thought that it would contain a significant residential population. It would however need to have a very high density to attain 100k and indeed the Nomination file puts it at 23708 - a density of 119.6

I haven't done a full trawl across all the European Capitals but, given their size and at European population densities, I can't see any of them having inscribed area populations anywhere near 100k. St Petersburg is the exception. This unwieldy nomination has (depending on definition) around 112 locations. Many of these are monuments, forts, palaces and parks but overwhelming them is the Historic Centre at a massive 3934ha (UNESCO hasn't even calculated the grand total!). Ok, it includes its share of monumental buildings, parks etc but also large number of residential streets. To its population need to be added those living in parts of a number of included surrounding towns - Kronstadt, Pushkin, Pavlovsk, Gatchina, Peterhof and Lomonosov. The central area alone only needs a population density of c 25 (way below that of central Vienna, Prague and Amsterdam) to exceed 100k. But proving/demonstrating this is proving quite difficult given Russian information sources. Am still working on it!

PS @Els - have you ever thought of adding WHS Area attibutes to the dB - unfortunately these are not searchable or sortable on UNESCO as far as i can see. Wouldn't need to be completed immediately - could be filled in longer term. Could have Core Area, Land and Sea division (where available), and Buffer.

No. It seems a lot of work to only be able to present a list of WHS ranked by size; I can't see any other use at the moment (do you?).
The UNESCO website has the data, maybe they will publicize it in the future on their Statistics page.

That is like all data provison - it isn't the full list which is important but the analysis the data permits, But no big deal

The Amsterdam Canal Ring is a very expensive area to live in, so the canal houses are mostly used by law firms and affluent elderly who live there. Nothing like the Ahmedabad pole, packed with large families!

It seems that it can be found in the official data dump - probably only the core zone in ha.

For Bamiyan and El Jam it has for example:
230 208 Rev Cultural Landscape and Archaeological Remains of the Bamiyan Valley 158,9265
234 211 Rev Minaret and Archaeological Remains of Jam 70

I guess this may be useful already? I could upload this easily.

Indeed it is the core zone area (as recorded on UNESCO with all its caveats and weaknesses!)- I wasn't aware of the existence of that data set. I solves my need (e.g questions like how big are the core areas of the inscribed "Historic Centres" of capital cities.......are there any Cultural sites with surprisingly large/small core areas........)
Thanks.

PS What is the difference between "unique_number" and "Id_no"? is it always 1 for 1? I thought it might be to do with multi country sites but i see that these are only in once each as a transboundary site and still have both numbers. so this list cant be used .e.g to count sites by country at least inso far as any transboundary sites are concerned.

I noticed that as well. It seems that UNESCO has done the same in their database as I did - a unique site has both an internal id (here 'unique number') and an official id (here the 'Id_no'). I started it because in the beginning the official id's were messy, and the internal id's in the database are just an auto_increment and good for quick queries. You never change the internal id, but the other one can (when a site merges for example, or you use a dummy before a site has been inscribed and has not been allocated its official id (before evaluation))

I note also that 95 sites have a core area of "0" and another 42 one of "space"......UNESCO obviously doesn't try to improve the accuracy of its data!

That doesn't cover ones where it is obvious that the recorded inscribed area is incorrect - e.g St Petersburg

We can try if that data can be found in the documentation.

I am currently trying to get the dataset of unesco regarding size in our db. Almost there!

UPDATE: here it is https://www.worldheritagesite.org/list/size