Most widely used projections

[PeteD]

Does anyone know of any kind of survey on the frequency of use of different projections?

In the meantime, in the absence of such a survey, I'd say that for world maps, the following projections seem to be the most widely used:

• Winkel tripel
• Robinson
• van der Grinten I (truncated and with repetitions)
• Mercator
• cylindrical equal-area
• other cylindrical projections that lie somewhere between the Mercator and the equirectangular – there are several that all look very similar

Are there any others that are as widely used as these? Kavrayskiy VII? Eckert IV? Hammer? Mollweide? Goode homolosine? I know that other projections like BSAM and the Ginzburg projections were used in Soviet atlases, but are they still widely used today?

[daan]

As you can imagine, there are lots of different ways to measure this, and I doubt they’d all turn up the same result. That said, I’m not aware of any research using any metric. I was involved in an effort some years ago to characterize atlas maps of the world throughout modern history (= 1900s+), but we never brought that paper to fruition, and that would be too narrow a survey for this purpose in any case.

If “frequency of use” means “viewed by anyone, each view, for any purpose, presently”, we have to suppose Web Mercator dominates overwhelmingly. Hundreds of millions views are served up each day on the major online wayfinding services. If we exclude online wayfinding services, then we have the fact that most countries publish a large-scale survey (1:5,000 — 1:50,000?) series of their territories, and those are overwhelmingly dominated by some parameterization or implementation of the ellipsoidal transverse Mercator.

Then there are regional maps, and those would be dominated by conic projections or cylindric (usually equal-area or conformal), depending on distance from equator. It’s hard to say what their frequency of views or frequency of production would be vis-à-vis all maps.

Your list starts looking possible as far as whole-world maps go, again excepting Web Mercator, which gets lots of views just because it’s right there on the phone whenever you want it. I might add Mollweide to the list, which appears frequently in many scientific domains. I’ll probably think of some others.

Cheers,
— daan

[PeteD]

Thanks for your reply!

If “frequency of use” means “viewed by anyone, each view, for any purpose, presently”, ...

I actually meant frequency of use by publishers rather than by viewers. "Frequency of production" would be a better term. I was thinking in terms of how many times a decision is made to use a particular projection rather than what projections viewers are most exposed to, although that would also be an interesting topic.

... we have to suppose Web Mercator dominates overwhelmingly. Hundreds of millions views are served up each day on the major online wayfinding services.

If we stick to frequency of production, then Web Mercator only gets one count per provider rather than one count per view.

Your list starts looking possible as far as whole-world maps go, again excepting Web Mercator, which gets lots of views just because it’s right there on the phone whenever you want it.

The "Mercator" on the list was intended to include Web Mercator. Would you regard them as fundamentally different projections?

How many of those views are of the whole world? I don't think of Internet maps as being world maps because viewers nearly always use them zoomed in.

I might add Mollweide to the list, which appears frequently in many scientific domains.

OK, another one to add to the list. I know it's often used for mapping the celestial sphere, including images of the cosmic microwave background, but didn't realize it was widely used in science in general.

[Atarimaster]

I know it's often used for mapping the celestial sphere, including images of the cosmic microwave background, but didn't realize it was widely used in science in general.

There seems to be a law saying that maps showing continental drift have to use a projection that shows the world in an ellipse… So sometimes it is Mollweide, sometimes Hammer, and sometimes it’s hard to tell because there are no graticule lines and today's state is missing.
In the brand-new edition of the Diercke Weltatlas, Mollweide is used for a fistful of World Ocean Maps, while almost all other world maps are Winkel Tripel. I don't know why they thought Mollweide would be particularly suitable for this application. Probably they didn’t read my blogpost about the Spilhaus maps.

I’d add Goode homolosine to your list – I don’t think that it is used very often, but IF there’s an “interrupted” world map, it’s always this one.
Apart from that, I couldn't name another widely used projection. Regrettably.

Kind regards,
Tobias

[Milo]

What counts as a "use" for a world map?

Actual world map pages in an atlas, designed to give an accurate, informative overview of the world? World map posters, designed to look cool on a wall more than anything? Maps in news stories and scientific papers, meant to illustrate some particular dataset rather than just the world itself? Video game boards, where the map is part of the gameplay rather than a static non-interactive setpiece? World maps on map projection list sites, which serve simply to give an example of what every given map projection looks like?

If you're counting all of them, and going strictly by a "number of publications" count regardless of viewership, then I think the news stories and scientific papers would probably dominate, because those are frequent short publications that will always have new reasons to keep grabbing maps to put in their articles, even if they're mostly lazy and unimaginative and keep using the same ones.

• Mercator
• cylindrical equal-area
• other cylindrical projections that lie somewhere between the Mercator and the equirectangular – there are several that all look very similar

By "between", and listing Mercator separately and explicitly, are you implying that you do not consider the equirectangular projection itself a common projection?

Because I do think it's a pretty common choice when people don't want to think about it too hard and are just in a rush to pick something quick and easy to use, especially when people want a rectangular map (whether for aesthetic reasons or because they actually need the cylindrical wraparound effect).

I'm under the impression equirectangular projections are more common than the cylindrical equal-area projections, really. People who understand (and care about) cartography well enough to see the value of an equal-area projection usually also realize the limitations of sticking to cylindrical ones. I don't think I've seen cylindrical equal-area projections much outside the context of cartography discussions about equal-area projections. Equirectangular projections appeal to people looking to make a simple, natural-looking compromise projection rather than optimizing any one variable.

I might add Mollweide to the list, which appears frequently in many scientific domains.

OK, another one to add to the list. I know it's often used for mapping the celestial sphere, including images of the cosmic microwave background, but didn't realize it was widely used in science in general.

There seems to be a law saying that maps showing continental drift have to use a projection that shows the world in an ellipse… So sometimes it is Mollweide, sometimes Hammer, and sometimes it’s hard to tell because there are no graticule lines and today's state is missing.

Equal-area projections in general are favored in statistics, particularly when mapping thing-per-area metrics like population density. Mollweide also seems to be the norm in meteorology/climatology and such. I've seen at least one case of an article using Mollweide, apparently as a scientific-community-accepted "default" map projection, in the one context that I really wouldn't recommend its use for: a climate simulation of weather on a tidally-locked planet, where, of course, the effect of latitude relative to the equator is far less dominant than on normal planets!

I’d add Goode homolosine to your list – I don’t think that it is used very often, but IF there’s an “interrupted” world map, it’s always this one.

That is a map that I've rarely seen outside of "you can do this, isn't it cool?" demonstrations of different world map projections. The fact that interruptions help to lower distortion is illustrative and helps to enlighten people on the basic tradeoffs that go into cartography, but has anyone ever used it outside of the context of talking about map projections?

On the other hand, it almost always shows up in layman-oriented galleries of different world map projections, even relatively small ones that don't have room for dozens of options.

[Atarimaster]

Because I do think it's a pretty common choice when people don't want to think about it too hard and are just in a rush to pick something quick and easy to use, especially when people want a rectangular map (whether for aesthetic reasons or because they actually need the cylindrical wraparound effect).

I'm under the impression equirectangular projections are more common than the cylindrical equal-area projections, really.

Hmmm.
I’ve seen Gall-Peters a couple of times, even outside of cartography discussions (probably no other cylindrical equal-area projections). But I don’t remember to have seen equirectangular maps in actual use (i.e. other than being a resource for map makers).
Maybe I have seen some of them, especially on websites, but I don’t remember it.

but has anyone ever used it outside of the context of talking about map projections?

Now that you mention it – I don’t think so.
Well, it is used as a thematic map on the wall map you can see here, but presumably, the idea again was to show various map projections.

[quadibloc]

other cylindrical projections that lie somewhere between the Mercator and the equirectangular – there are several that all look very similar

It's true that I've seen the Gall Stereographic projection in use a few times, but I think the Miller Cylindrical projection is much more common, at least in the United States - and I can't think of any others in that category that are at all popular.

Thanks for your reply!

I might add Mollweide to the list, which appears frequently in many scientific domains.

OK, another one to add to the list. I know it's often used for mapping the celestial sphere, including images of the cosmic microwave background, but didn't realize it was widely used in science in general.

Actually, the Aitoff-Hammer projection of the sphere is far more often used than the Mollweide for mapping the celestial sphere. For the cosmic microwave background and other purposes. It's also used, particularly in oblique aspect, for world maps too.
But the interrupted Mollweide, and the interrupted Sinusoidal - and, of course, the Goode Homolosine projection - are used fairly often for world maps. Or, at least, they were; perhaps if one is only looking at atlases currently in print, the frequencies of projections would be different than what it seems to me, based on the experience of all the different atlases I've ever encountered.
The Aitoff - the one based on the Azimuthal Equidistant instead of the Azimuthal Equal-Area - is basically not used at all nowadays, but I've encountered the old French publications, by Aitoff himself, in which it was used; I've discounted those.
So I remember from back in the 'sixties the Eckert IV projection being used, as Hammond chose it as a nice compromise equal-area projection for a while.

And if one looks at the atlases of different countries, things change. So in Britain, one sees the projections Guy Bomford came up with a lot; in the People's Republic of China, they have a projection all their own for world maps which is derived from the one that the old Soviet Union used to use in all of its atlases.

[PeteD]

What counts as a "use" for a world map?

Actual world map pages in an atlas, designed to give an accurate, informative overview of the world? World map posters, designed to look cool on a wall more than anything? Maps in news stories and scientific papers, meant to illustrate some particular dataset rather than just the world itself?

Yes.

Video game boards, where the map is part of the gameplay rather than a static non-interactive setpiece?

Er ... um ... I don't know. I hadn't thought about this.

World maps on map projection list sites, which serve simply to give an example of what every given map projection looks like?

No, because there was no decision to choose a particular one of the projections over all the others.

By "between", and listing Mercator separately and explicitly, are you implying that you do not consider the equirectangular projection itself a common projection?

Yes, that's right, although I'm happy to be proved wrong about this.

I don't think I've seen cylindrical equal-area projections much outside the context of cartography discussions about equal-area projections.

I've seen a fair few Gall–Peters wall maps.

Equal-area projections in general are favored in statistics, particularly when mapping thing-per-area metrics like population density.

Well, for "thing-per-area metrics", equal-area projections should be used but aren't always, as you can see from a Google image search for population density maps. The Robinson projection seems quite popular, and some maps even use the Mercator.

If the data shown isn't a "thing-per-area metric", then I don't see any particular reason to use an equal-area projection.

I've seen at least one case of an article using Mollweide, apparently as a scientific-community-accepted "default" map projection, in the one context that I really wouldn't recommend its use for: a climate simulation of weather on a tidally-locked planet, where, of course, the effect of latitude relative to the equator is far less dominant than on normal planets!

What would you have used? A two-hemisphere projection?

The fact that interruptions help to lower distortion is illustrative and helps to enlighten people on the basic tradeoffs that go into cartography, but has anyone ever used it outside of the context of talking about map projections?

In a children's atlas that I had as a kid, on every page with a regional map, there was a small inset with a world map showing where in the world that region was located. Those small world maps used an interrupted projection*, although I couldn't say which one. That is, however, the only time I remember seeing an interrupted projection outside of the context of talking about map projections.

*I mean interrupted along more than just one meridian.

[PeteD]

It's true that I've seen the Gall Stereographic projection in use a few times, but I think the Miller Cylindrical projection is much more common, at least in the United States - and I can't think of any others in that category that are at all popular.

It's possible that all the ones that I've seen were either the Gall stereographic or the Miller cylindrical. You could show me any of the Gall stereographic, Miller cylindrical, Tobler cylindrical I or Urmayev cylindrical II and I wouldn't be able to tell you which one it was.

The Aitoff - the one based on the Azimuthal Equidistant instead of the Azimuthal Equal-Area - is basically not used at all nowadays, but I've encountered the old French publications, by Aitoff himself, in which it was used; I've discounted those.

Wasn't Aitoff Russian?

[quadibloc]

Wasn't Aitoff Russian?

Yes, but David Alexandrovitch Aitoff moved to France at the age of 23, and remained there for the rest of his life.