Wikidata:Property proposal/population density

population density

Originally proposed at Wikidata:Property proposal/Place

Not done

Description	density of the human population of a particular area/region.
Represents	population density (Q22856)
Data type	Quantity
Template parameter	population_density in `{{Geobox}}`
Domain	subclass of territory (Q1496967)
Allowed values	any positive number
Allowed units	we may need to create some units here
Example	Lisbon (Q597) → 6458 people/km²
See also	Discussion at https://www.wikidata.org/wiki/Wikidata:Project_chat#density_vs._population_density

Motivation

Looks like we do not have a property to describe population density, and people start adding the numbers under "density", which is wrong. We should add this property.

Support. If the intention is to restrict this to human populations then the units can just be "per km²" but if we want other populations then we'll need "people per km²". I support either. Thryduulf (talk) 20:59, 15 August 2016 (UTC)[reply]
Oppose Density is the result of population/land area. I don't see the neccessity to create a property when we can calculate it with other properties. --Fralambert (talk) 04:22, 16 August 2016 (UTC)[reply]
Oppose. This should be calculated. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:06, 16 August 2016 (UTC)[reply]
- It may be calculated, indeed, if the source data is known (not always so). Still, a lot of encyclopedic sources, reference literature and Wikipedia infoboxes produce these numbers. Judging by use of wrong "density" property, there's demand for this in Wikidata too. Also, there might be different population densities for parts of the whole, for example. --Laboramus (talk) 04:46, 19 August 2016 (UTC)[reply]
Support. Would be better to not restrict to people - many ecology studies have population densities of different species (or also of breeding pairs). This sort of detail could be added as a qualifier.
Population densities may be estimated from sampled areas rather than being simply being a matter of dividing the count by the area - an example of sampling methodology from British Trust for Ornithology (Q2925763).
Also, the given population density in a source may be given to a certain number of significant figures which the source has deemed appropriate, and not overly concise, for example the Office for National Statistics (Q1334971) gives population density from censuses as just 2 significant figures (example here). BTW - they give the figure as "per ha²"... Robevans123 (talk) 13:05, 18 August 2016 (UTC)[reply]
- That's a different property. The one proposed here is specifically framed in terms of human population density (or is otherwise very badly explained). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:00, 18 August 2016 (UTC)[reply]
  - Yes, I meant human population. --Laboramus (talk) 04:46, 19 August 2016 (UTC)[reply]
    - @Laboramus: Can you amend the proposal (either the name or description or both) to make this clear (and to avoid people using this property for other population densities)? Also amend the units to "per km²" as per Thryduulf's comment? Robevans123 (talk) 14:58, 19 August 2016 (UTC)[reply]

Yes - a human property density may be calculated, but this is fraught with difficulties:

to what precision should the result be displayed? Naively this may be derived from the number of significant figures of the population and the area, but many population figures given are themselves estimates, or have instrinsic unknowns (do we know how many households complete a census and how accurately)?
what if the population figure and the area data come from different sources and/or different times (boundaries can and do change)?
Published figures for population densities are often rounded to a whole number (or sometimes one decimal point). Presumably the Office for National Statistics (Q1334971) have a good idea of what constitutes an appropriate level of precision for the figure to be meaningful and useful.

Wikidata should ideally include verifiable information. If we have verifiable information from a reliable authoritative source, surely we should include that, rather than relying on, for example, a Wikidata enabled infobox, to carry out the calculation (with all the intrinsic problems of displaying appropriate precision that would entail)? BTW - I'd be quite happy for infobox code to carry out the calculation if no population density is available, but it should clearly indicate that it is a calculated value and not a sourced value... Robevans123 (talk) 15:36, 19 August 2016 (UTC)[reply]

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘

"to what precision should the result be displayed?" that's a matter for the consumer, not Wikidata. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:44, 19 August 2016 (UTC)[reply]

Speaking as a consumer, I'd like a well-sourced figure from a authorative source, thank you very much! Robevans123 (talk) 16:29, 19 August 2016 (UTC)[reply]

I'm sorry but, as explained, this would be a derived property, which is currently not supported in Wikidata. In the future, it would be a great idea to show derived values or to save internally some frequently-used derived values in order to increase query performance but, currently, this feature is not available.

Oppose. Anyway, many thanks for your feedback. --abi án 21:23, 20 August 2016 (UTC)[reply]

That's a mistaken claim. The property is not proposed as a derived property. It's quite possible that a source has data on this with higher precision than can be derived with public data. ChristianKl (talk) 13:18, 30 August 2016 (UTC)[reply]

It's not about displaying precision but about the inherent precision that the data has. I agree with Robevans123 that a consumer wants well sourced data and wants his data at the precision that a source provides. ChristianKl (talk) 13:13, 30 August 2016 (UTC)[reply]

Oppose. Redundant, and likely to lower data quality as a result. --Yair rand (talk) 10:32, 24 August 2016 (UTC)[reply]
Oppose sorry but I still fail to see the need and the usefulness of this property as the value can easily be calculated with two common properties (and BTW, precision is easily calculable too). Cdlt, VIGNERON (talk) 09:28, 29 August 2016 (UTC)[reply]
Support Different sources have different rules for calculating precision. If you devide a number with 3 degrees of accuracy by another number with 3 degrees of accuracy you get a number with less degrees of accuracy. If a source publishes population, area and population density all with 3 degrees of accuracy you can't reconstruct the population density that the original source has by doing math.

Given NPOV we should be able to document different view and there's no reason to limit the world in this case. When it comes to making Wikidata easy to use for list making, I also see advantages to having a straightfoward property. ChristianKl (talk) 13:09, 30 August 2016 (UTC)[reply]

Support--Mikey641 (talk) 02:46, 21 October 2016 (UTC)[reply]
I've marked this as ready as it seems to me (although I am neither a property creator nor uninvolved here) that the arguments in opposition to the property have all been addressed - namely that this is not always calculable (a) at all because the source data for the population and/or area concerned is not always known; (b) to the same level of precision if only post-rounding figures are given or the source figures are given to different precisions. I expect a property creator to fully read the discussion and make a determination for themselves though - the "ready" flag is intended as a signal for them to do so. Thryduulf (talk) 14:45, 6 November 2016 (UTC)[reply]
Oppose redundant. --Gloumouth1 (talk) 14:39, 17 November 2016 (UTC)[reply]
- @Gloumouth1: Why do you think it is redundant in all cases? Please respond to the arguments made above that it is not always calculable - either at all or to the same level of precision. If calculability is not the basis for your opposition, please explain your reasoning further. Thryduulf (talk) 11:56, 18 November 2016 (UTC)[reply]
Oppose like say before, it a calculation of 2 parameter that we have. - yona b (talk) 08:05, 28 December 2016 (UTC)[reply]

Not done For most cases it can be calculated easily and it doesn't need to be stored (and maintained) in WD. For estimated population densities or different uses please propose a more specific property. @Laboramus, Thryduulf, Fralambert, Pigsonthewing, Robevans123: @Abián, ChristianKl, Yair rand, VIGNERON, Mikey641: @Thryduulf, יונה בנדלאק:--Micru (talk) 15:54, 6 January 2017 (UTC)[reply]