Other religions’ sites — how this layer is made

What you are looking at

Each point is a religious site belonging to a religion that has no map of its own on this site — recorded in OpenStreetMap, and covering some forty traditions between them. Chinese folk religion, Confucianism, Vietnamese folk religion, spiritualism, Tenrikyo, Cao Dai, Bahá'í, Zoroastrianism, Yazidi, Druze, Afro-Brazilian traditions and a long tail beyond. Around 9,700 worldwide, coloured by which religion it is and sized by the area of its footprint.

The nine religions with their own maps — Christian, Muslim, Buddhist, Hindu, Shinto, Jewish, Taoist, Sikh and Jain — are not on this one. This is everything else.

A residual that turned out to be the least distorted map of the family

Every other map in this set spends a section explaining why its distribution is not what it appears to be. This one has the opposite problem, and it is a nice one to have.

Its top country holds 15.6%. Compare Shinto at 99.5% and Taoist at 58.4%. Six countries on four continents each hold above 4%. A map assembled from what was left over turns out to be the most globally spread thing in the family — because it is not one religion with one heartland, but forty religions whose heartlands are all over the world, and no single mapping community's habits dominate it.

That said, China, Vietnam and Taiwan together hold a large share, and the reason is visible in the colours: the three largest members — Chinese folk religion, Confucianism and Vietnamese folk religion — are East and South-East Asian folk and ancestral traditions, and they are 51.5% of the map between them.

Colour shows the religion

The other nine maps colour a dot by what kind of site it is — a temple, a cemetery, a school. That is the right choice there, where the religion is the whole map and the kind of site is the only thing left to distinguish. Here it would be the wrong one: this map holds forty traditions, and which of them a dot belongs to is the first thing worth knowing about it.

So colour shows the religion, and you can switch any of them off in the legend. Eleven are named, with their share of the map:

Chinese folk religion — 20.1%, much the largest
Confucian — 12.1%
Vietnamese folk religion — 11.1%
Spiritualist — 7.7%
Tenrikyo — 5.3%, a Japanese new religion
Vodou — 3.3%
Pagan — 2.9%
Bahá'í — 2.5%
Benzhu — 2.5%, the local-deity cults of the Bai people of Yunnan
Cao Dai — 2.2%, in yellow, which is its own colour
Zoroastrian — 0.9%

The kind of site is still there. Click any dot and the panel names it — place of worship, burial site, grounds, school, monastery — along with the exact religion as OpenStreetMap records it. Size still reflects the kind of site too, as described below.

The grey class, and an argument this page is having with itself

The remaining 28.6% is drawn in grey and labelled "other religion". That is a lot of grey, and there is no version of this map where it is much less.

It holds animist (153 sites), ancestor veneration (102), Scientologist (99), Konkokyo (94), shamanic (58), Afro-Brazilian matriz africana (55), Yazidi (43), Umbanda (43), Antoinist (38), Yiguandao (37), Yarsan (25), Seicho-no-Ie (22), Candomblé (20), Druze (20) and a great many more.

"A great many more" is the honest part. There are 1,377 distinct religion values on this map and 1,093 of them have exactly one site. Only 71 have five or more. This map is the place every religion goes when no other map claimed it, so it collects every local name, every alternative spelling and every value written in a script the tagger's neighbours use — and one fifth of the whole map sits below the thirtieth most common value. No legend, at any length, reaches into that.

This sits awkwardly beside something else this page says, and it would be dishonest to leave the tension unstated. The section below explains that small religions are deliberately not folded into their large neighbours — that Yazidi and Druze sites are kept out of the Muslim map on the principle that absorbing a small religion into a big one is precisely how it disappears from maps. And here they are, in a grey bucket.

The difference is real, and it is this. An alias would change what the data says: a Yazidi site would become a Muslim one, and no amount of clicking would recover it. This grouping changes only what the legend has room for. The site is still recorded as Yazidi, the panel still says Yazidi, and the grouping lives in this website's display code rather than in the map data, so it can change tomorrow without rebuilding anything.

That is a genuine distinction rather than a comfortable one, and there is a limit behind it: a legend of a thousand rows is not a legend. Eleven named classes plus grey is where legibility and completeness stopped trading well against each other — going to fifteen would recover four percentage points and cost four more colours that a colour-blind reader could not tell apart. The floor is about 20% regardless, because that is the share of this map made of traditions with a handful of sites apiece.

What is deliberately not here

Jainism is not here; it has a map of its own. At 738 sites it would be the fifth-largest member of this one. That is not a size decision — four traditions on this map are larger — but the same judgement that gives Jainism a class of its own on the places of worship map: a named world religion earns its own map where a folk grouping does not.

Sites tagged for several religions at once are excluded. About 1,700 sites are recorded as `multifaith`. Including them would have made "several religions at once" one of the largest members of a map of distinct religions, which would be describing the map's own construction rather than the world. The places of worship map keeps a multifaith class, because there the question being asked makes it a real answer.

Small religions are not folded into large ones. Yazidi, Yarsan and Druze sites stay here rather than being absorbed into the Muslim map; Confucian, Chinese folk and Vietnamese folk sites stay here rather than the Taoist map. Grouping small religions under their nearest large neighbour is exactly the mechanism by which they disappear from maps, and this one does not do it. See the section above for how that squares with the grey class.

A few hundred sites are tagged with a non-answer — `none`, `yes`, `unknown` — and are excluded, since those say nothing about a religion.

Unitarian Universalist sites are not on any map here, and that was a difficult decision. There are 672 of them. Unitarian Universalism draws on Christianity and on a good deal besides, and it does not ask its members to hold any one of those things; putting its meeting houses on a map of Christian sites would tell a reader it is purely a Christian religion, which is not what it says about itself. Leaving them off is the lesser of two misrepresentations rather than a good outcome, and the honest thing is to say so here rather than let 672 sites go quietly missing.

One site on this map is almost certainly wrong

The largest site here is 34 square kilometres, tagged Bahá'í, in Kenya — and its own name says it is a Later Stone Age archaeological area. A Bahá'í holy place of that size does not exist.

It is an error in the source data rather than in this map's rules, which did exactly what the tags asked. It matters more here than the same error would elsewhere, because symbols are sized by area and this is a small map: one mis-tagged polygon is the biggest mark on the world view.

Colouring by religion has made it easier to spot rather than harder — a vast blue Bahá'í polygon in Kenya, with no other Bahá'í site anywhere near it, now looks wrong at a glance in a way that a generic grounds colour never did.

For contrast, the next largest are real: Dragon Springs in New York (Falun Gong) at 1.45 km² and the Radha Soami Satsang Beas campus at 1.24 km².

Anyone can correct OpenStreetMap, and this map will follow when someone does.

Size

Each symbol is sized by footprint area — the ground the mapped outline covers, in square metres measured on the globe. Symbol area is proportional to footprint, so a site four times the area draws twice as wide.

About 37% of these sites are mapped as a single point with no outline, and are given a nominal, typical size rather than the smallest one. Because this is a sparse map — under ten thousand sites across the whole world — the symbols also have a higher floor than the larger maps here, or most of it would be invisible at the zooms people actually use.

Places of worship are drawn as though their footprint were twice its real size — about 1.4 times the width. This is a deliberate weighting, and the only one on the map. It matters more here than anywhere else in the family: 15% of this map is burial grounds, the highest share but one, because Bahá'í, Yazidi and Zoroastrian sites in OpenStreetMap are disproportionately cemeteries and dakhmas, and Chinese folk religion contributes very large cemetery complexes across Malaysia and Singapore. Without the weighting, a map of forty religions would be mostly a map of their graveyards.

Sites with no name

Around 79% of the sites here carry a name — the highest share of the family — and the unnamed ones are shown just the same.

A picture of the present

Today's map, not a history. OpenStreetMap is edited continuously; this copy is refreshed periodically. There is no year control, and a site closed or converted long ago may still be mapped until a mapper updates it.

Source and licence

Back to the map