Hindu sites — how this layer is made

What you are looking at

Each point is a Hindu site of any kind recorded in OpenStreetMap — mandirs and the grounds they stand in, but also cremation grounds, Hindu schools, wayside shrines, mathas, offices and community buildings. Just over 57,000 worldwide, drawn as a dot sized by footprint area and coloured by what kind of site it is. At the closest zooms the dot is joined by the building's mapped outline, where one has been drawn.

This is the companion to the places of worship map, which colours every religion's places of worship by religion. This one takes a single religion and shows everything of it that is mapped.

Click any site to see what OpenStreetMap records about it — the kind of site it is, and its denomination where that was recorded. Where the site has a Wikipedia article, a short summary is fetched live and shown in your own language; and every site links to the OpenStreetMap object it came from.

Where the map is right, and where it is wrong

This is the one map in this set whose headline shape needs no apology. India holds 71% of these sites and Nepal 12% — 83% between them — and that is genuinely where Hindus live. On most of this site's religion maps a concentration like that would be telling you where the mappers are; here the mapping and the reality point the same way.

The distortion on this map is at the other end of the table, and it is an under-count:

Mauritius has 123 mapped sites. About half the island's 1.26 million people are Hindu and it has many hundreds of temples.
South Africa has 57. It has a Hindu population of well over a million.
The United Kingdom has 191, which is plausibly close to the true number of British mandirs.

So the gap between the UK and Mauritius on this map is a difference in mapping effort, not in Hindu presence. Where a country's diaspora sits in a well-mapped place, its temples are nearly all here; where it does not, they are largely missing.

And India's own 42,000 is simultaneously the largest count on the map and, in absolute terms, almost certainly the most incomplete. There is no independent count of Indian temples to measure that against, so the size of the gap is not something this page can tell you.

Colour

Colour shows what kind of site it is, not which religion — the religion is the whole map. All ten kinds appear:

Place of worship — the mandir itself. 94.7% of this map, the highest share of any religion map here.
Grounds — the plot mapped as its own polygon rather than the building.
Marker — a wayside shrine rather than a building anyone enters.
Burial site — see below; this row is not what its name suggests.
School, former site, religious building, office, monastery, community building — each present, in hundreds or fewer.

The burial row deserves an explanation, because the number looks wrong for Hinduism. Hindu funerary practice is cremation, so a count of 408 burial sites is surprising until you see what is in it — and 31 of them are crematoria. A cremation ground is a first-class Hindu religious site, and it had nowhere else to go: these nine maps share one vocabulary of site types so that you can compare them, and that vocabulary has a burial row and no cremation row. Renaming the row for this map alone would break the comparison the shared classes exist to allow, so it is explained here instead.

The monastery row is real too: `amenity=monastery` reaches the matha, a living Hindu institution rather than a borrowed Western category.

Which sites count as Hindu

A site is on this map when OpenStreetMap's `religion` tag says `hindu`. That tag is what the mapper stated, and it is taken at its word — a building's architecture and its name are not second-guessed.

Jain sites are not folded in. Jainism is a distinct religion rather than a branch of Hinduism, and it has a map of its own. The same reasoning keeps Sikh sites on theirs.

Sites tagged with a branch of Hinduism instead of the religion itself — `vaishnavism`, `shaivism` — are folded in where they occur, but this turns out to matter almost not at all: between them they are ten sites on the whole map.

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 52% of these sites are mapped as a single point with no outline, and so have no footprint. Rather than draw them all at the smallest size, they are given a nominal, typical size. Whether a site is mapped as a point or an outline is mostly a matter of how thoroughly it was mapped, not of how large it stands.

Mandirs 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 less here than on the other religion maps, because at 95% places of worship almost everything on this map is weighted; its effect is mostly to lift the temples against the schools and cremation grounds among them.

Footprint is physical extent, not importance or congregation. A large temple complex covers more ground than a revered small one, and the map says nothing about which matters more to the people who use it.

Sites with no name

77% of the sites here carry a name, and the unnamed ones are shown just the same. A site with no name is one nobody has recorded a name for, not one that is not there.

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