Care Abroad

a seriesOpenStreetMap hospitals, clinics and dental surgeries

also: OSM facilities harvest · Overpass hub query

Mappers call them POIs. Each row is a point somebody typed a name onto, and the tags are whatever that person thought worth typing.

Every named hospital, clinic and dental surgery OpenStreetMap carried within a set radius of each hub on this site, fetched through the Overpass API. The copy on disk holds 14,356 rows across 14 hubs, fetched on 16 and 17 September 2026: 5,532 tagged amenity=clinic, 5,323 amenity=hospital, 2,651 amenity=dentist, and 788 carrying a healthcare tag with no amenity tag at all.

The story Harvested · osm

The count follows the mappers, not the medicine. Bengaluru returns 2,548 rows inside 25 km and Bangkok 521 inside the same 25 km; Delhi NCR returns 2,275 inside 40 km, Mumbai 1,849, Chennai 1,529, Istanbul 1,297 inside 30 km, Kuala Lumpur 1,287, Bogotá 894, Barcelona 621, Budapest 549, Dubai 545, İzmir 206, Amman 139 and Antalya 96 inside 25 km. India supplies 8,201 of the 14,356 rows. Nobody should read that as India having sixteen times the clinics of Thailand; India has a larger and more active mapping community, and Thai medical addresses are often mapped in Thai script under names a Latin-script query does not reach.

The tags thin out fast. Of 14,356 rows, 1,621 carry a website, 1,581 a telephone number, 3,775 a street address, 1,375 an operator — and 42 carry a bed count. A bed count on three rows in a thousand is not a capacity dataset.

Inference — the country code on each row shows how the sausage is made. It is taken from the mapper's own addr:country tag where there is one, and from the hub's country otherwise. One clinic in Dubai Marina carries addr:country=UA, so the harvest files it under Ukraine. One row in 14,356, visible only because somebody counted; the same class of error is invisible everywhere it happens to agree with the hub.

How it works Harvested · osm

tools/harvest_osm.py runs one Overpass query per hub at the radius the hub record names: nwr["amenity"="hospital"], amenity=clinic, amenity=dentist and healthcare in hospital, clinic, centre, dentist, fertility_clinic, rehabilitation or laboratory, each requiring a name, around the hub's own point. Ways and relations are reduced to their centre. Rows are deduplicated by OSM id across hubs, and twenty-eight tag keys are kept — name, operator, healthcare speciality, beds, website, phone, address parts, wheelchair, wikidata and a few more.

Notes

ODbL 1.0, which is share-alike: the attribution travels with the rows onto the page and into anything built from them. © OpenStreetMap contributors.

The particulars

publisherOpenStreetMap
regionEverywhere
publisherOpenStreetMap contributors, through the Overpass API
unitnamed points, one row per OSM object
yearsa snapshot, fetched 16 and 17 September 2026
countries9
licenseODbL 1.0
updated by publishercontinuously — OpenStreetMap changes by the minute, and this file does not
countsPoints, ways and relations inside the radius that carry a name and one of the queried amenity or healthcare tags on the fetch date.
does not countAnything nobody mapped. A hospital missing from this file is missing from OpenStreetMap, which is a fact about mappers and not about the city: absence here is absence of mapping. The rows carry no accreditation, no ownership, no price, no quality measure, no patient volume and no indication that a place treats foreigners — and a clinic that closed last year sits in the file until somebody edits it out. A name is not a licence: a row proves only that a point on a map once had that name typed on it.
linksOpenStreetMap · ODbL 1.0
confidencehigh · updated 2026-09-17

What it connects to

a clusterBangkok521 rows inside 25 km of the hub's point, against Bengaluru's 2,548 inside the same radius.and of this page it says: The harvest that pulls every mapped hospital, clinic and dental surgery within 25 kilometres of this point.a clusterBengaluruThe largest harvest on this site at 2,548 rows, and a lesson in reading map coverage as medical density.and of this page it says: The harvest that pulls every mapped hospital, clinic and dental surgery within 25 kilometres of this point.a clusterIstanbul1,297 rows inside 30 km, in the city with the loudest advertising in this trade.a seriesThe JCI list of accredited organizationsThe opposite selection: a register of who applied and paid, against a map of who got mapped.and of this page it says: The opposite selection problem: a map of who got mapped, against a register of who applied.a long oneWhat the data cannot seeAbsence of mapping, read as absence of a hospital, is the commonest mistake this file invites.and of this page it says: Where absence is absence of mapping, and a row proves only that somebody typed a name.a long oneHow this was builtThe Overpass query, the radii and the tag list are all in the method.and of this page it says: The harvest that supplies the hub maps, under ODbL with attribution.

Pages that point here

a cluster · on this pageChennaiThe harvest that pulls every mapped hospital, clinic and dental surgery within 20 kilometres of this point.a cluster · on this pageColomboThe harvest that pulls every mapped hospital, clinic and dental surgery within 15 kilometres of this point.a cluster · on this pageDelhi NCRThe harvest that pulls every mapped hospital, clinic and dental surgery within 40 kilometres of this point.a cluster · on this pageKuala LumpurThe harvest that pulls every mapped hospital, clinic and dental surgery within 25 kilometres of this point.a cluster · on this pageMumbaiThe harvest that pulls every mapped hospital, clinic and dental surgery within 25 kilometres of this point.a cluster · on this pagePenangThe harvest that pulls every mapped hospital, clinic and dental surgery within 15 kilometres of this point.a cluster · on this pagePhuketThe harvest that pulls every mapped hospital, clinic and dental surgery within 15 kilometres of this point.a cluster · on this pageSeoulThe harvest that pulls every mapped hospital, clinic and dental surgery within 20 kilometres of this point.a cluster · on this pageSingaporeThe harvest that pulls every mapped hospital, clinic and dental surgery within 15 kilometres of this point.a cluster · on this pageTaipeiThe harvest that pulls every mapped hospital, clinic and dental surgery within 20 kilometres of this point.a series · on this pageNatural Earth country outlinesThe other map layer: this one draws countries, that one draws the clinics inside nine of them.

Sources

Where it came from: Cited a source named here · Harvested pulled from an open dataset · Trade practice how the trade works, hedged · Inference this project's own reasoning · Field somebody stood there. This record as JSON.

Pass it on
BlueskyMastodonXFacebookRedditWhatsAppEmail