English | Français | العربية

@geoalgeria/cliniques

Algeria's clinics and proximity-care facilities, as data you can install.

1,917 geocoded care facilities across 66 wilayas of Algeria, every one with coordinates, classified by type (polyclinique · salle de soins · centre de santé · maternité · clinique), most with Arabic and/or French names, and commune/wilaya linkage. Extracted from OpenStreetMap. This is the community tier of the health sector: the Ministry of Health registry (CHU/EPH/EHS/EPSP) is deliberately excluded here and lives in @geoalgeria/sante. Shipped as JSON, CSV, GeoJSON, and TypeScript. Part of GeoAlgeria.

npm install @geoalgeria/cliniques

import cliniques from "@geoalgeria/cliniques";

const all = cliniques.cliniques();   // 1,917 geocoded care facilities

// The public proximity tier of one wilaya
const proximite = cliniques.cliniquesByWilaya("16")
  .filter((c) => c.type === "polyclinique" || c.type === "salle_de_soins");

// Facilities that tag an emergency service
const urgences = all.filter((c) => c.emergency);

What you can build

• "Care near me" locators, coordinates on all 1,917 records, ready for a map or nearest-facility distance sorting.
• Proximity-care coverage maps, count polycliniques and salles de soins per commune or wilaya, the structures Algerians actually walk into first.
• Bilingual directories, thousands of Arabic and French names side by side, with phone, opening hours and speciality where the map carries them.

What's inside

| Dataset | Count | Coordinates | Notes | | --- | --- | --- | --- | | Care facilities | 1,917 | ✅ all | 1,610 named, 66 wilayas |

By type

| Type | Count | Meaning | | --- | --- | --- | | clinique | 1,144 | clinic (عيادة / مصحة), mostly private practice-level care | | polyclinique | 406 | polyclinique (عيادة متعددة الخدمات), public proximity tier | | salle_de_soins | 199 | salle de soins / dispensaire (قاعة علاج / مستوصف) | | centre_sante | 142 | centre de santé / centre de soins (مركز صحي) | | maternite | 26 | maternité / clinique d'accouchement (مصحة توليد) |

This is an OpenStreetMap extract, not an official registry. Coverage is partial and uneven by wilaya, and three wilayas (54 In Guezzam, 62 Bir El Ater, 63 El Aricha) carry no mapped facility at all. Nothing official enumerates this population, so the package ships no coverage percentage: the Ministry of Health publishes counts for the registry tier this package excludes, and no public register lists private clinics. Counts move as OpenStreetMap is edited; each rebuild reflects the current state of the map.

It never republishes a hospital @geoalgeria/sante already ships, and the two must not be summed. sante is the registry tier: 668 public establishments (CHU, EPH, EHS, EPSP) from the Ministry of Health. This package is the community tier. 121 of sante's records reference an OSM element by id. The 86 of them on a hospital (CHU, EPH, EHS) are excluded here by construction, so no hospital is published twice under the same OSM element. The other 35 are EPSP records, whose element anchors the entity on one of the proximity facilities it runs, and that facility stays here: entity in sante, facility here. Be precise about what that does and does not guarantee: sante's other 547 records carry no OSM reference at all, so the same physical establishment can still appear in both packages, under different coordinates and different ids, with nothing mechanical to detect it. The two describe different tiers of a health system, so adding 668 to 1,917 counts nothing real.

Type is inferred from the name. A polyclinique names itself polyclinique/عيادة متعددة الخدمات, a salle de soins قاعة علاج/مستوصف/dispensaire, a centre de santé مركز صحي/centre de soins. Order matters: the facility words are matched before the bare word "hôpital"/مستشفى, because Algerian mappers use that word for proximity structures too (10 records name themselves both ways, e.g. "Polyclinique des consultations spécialisées" tagged name:ar=مستشفى بودغن). The plain word "clinique"/عيادة/مصحة counts as a facility word too, one rank below the three specific types. Everything left over is clinique, including the 307 unnamed clinic-tagged points, which the tag alone already identifies as care facilities.

What was excluded, and why. The pull returns 2,962 OSM elements; 979 are dropped before anything is emitted:

| Excluded | Count | Reason | | --- | --- | --- | | hopital | 365 | hôpital / مستشفى / المؤسسة الاستشفائية / EPH / EHS / EHU / centre anti-cancer, the registry tier (@geoalgeria/sante) | | unnamed_hospital | 249 | no name at all and tagged as a hospital, so it cannot be told apart from the registry tier | | sante_overlap | 85 | the OSM element is one a @geoalgeria/sante hospital-tier record (CHU/EPH/EHS) already ships, whatever it is named here. Elements referenced by a sante EPSP record are not excluded: there the reference is a geocoding anchor on the entity's seat and the element is usually a facility this package should carry | | cabinet | 96 | single-practitioner practice: the word cabinet, or a name that is just a practitioner (Dr X, الطبيب …) | | epsp_entity | 105 | the EPSP administrative entity itself (its facilities stay) | | hospital_subfeature | 54 | part of a hospital mapped as its own point: an entrance, a ward, "Service de radiologie", a bare "urgences". When the name also says hospital the record is reported as hopital instead | | pharmacie | 6 | pharmacy, belongs to @geoalgeria/pharmacies | | chu | 14 | centre hospitalo-universitaire | | institut_pasteur | 3 | research institute rather than a care facility | | paramedical | 2 | paramedical training school, education rather than care |

Hospitals are queried on purpose even though none ship: they are the only way to reach the clinic-class records Algerian mappers file under amenity=hospital.

Private hospital establishments are kept, on purpose. An "EHP" or "établissement hospitalier privé" is a clinique privée, which is exactly this package's population, so the privacy check runs before every registry pattern and those records come in as clinique with sector: "private". The mirror-image abbreviations are the reason the order matters: EHP (privé) must never be read as EPH (public).

Sector is asserted only on signal. public when OSM says operator:type (including university, an EHU being a public teaching operator), or structurally for polyclinique and salle_de_soins (both are public structures of the Algerian proximity-care system by definition); private on operator:type=private or a privé/خاصة name read across every name tag, since a record can carry its only ownership signal in name:en. 616 records are public, 67 private, and the remaining 1,234 stay null. Most cliniques are private in practice, but the map does not say so, so the field does not pretend to know.

Also on each record: speciality (from OSM healthcare:speciality, on 166 records), address (from addr:* tags, on 654), phone (on 116), opening_hours (on 168) and emergency (true on the 69 records tagged emergency=yes, never false: a silent map is not a claim that there is no emergency service).

Formats

The npm package ships the JSON (importable directly):

import cliniques from "@geoalgeria/cliniques/data/cliniques.json" with { type: "json" };
// or via CDN, no install:
// https://cdn.jsdelivr.net/npm/@geoalgeria/cliniques/data/cliniques.json

The loaders and record shapes are fully typed, TypeScript definitions ship in the package:

import cliniques, { type Clinique } from "@geoalgeria/cliniques";
const all: Clinique[] = cliniques.cliniques();

CSV and GeoJSON are in the repo under data/ and bundled in every GitHub Release:

data/
  cliniques.json              # 1,917 care facilities (array)
  metadata.json               # sources, counts, coverage, updated
  csv/cliniques.csv           # repo + Release bundle (not in npm tarball)
  geojson/cliniques.geojson   # Point features

Record shape

{
  "id": "02-00006",
  "name": "عيادة متعددة الخدمات",
  "name_fr": null,
  "name_ar": "عيادة متعددة الخدمات",
  "wilaya_code": "02",
  "commune_code": "0207",
  "commune": "Beni Haoua",
  "lat": 36.530021,
  "lng": 1.58226,
  "geo_precision": "exact",
  "geo_method": "osm_node",
  "source": "osm",
  "refs": {
    "osm": "node/4144869592"
  },
  "type": "polyclinique",
  "type_label_fr": "Polyclinique",
  "type_label_ar": "عيادة متعددة الخدمات",
  "sector": "public",
  "speciality": null,
  "address": null,
  "phone": "027753479",
  "opening_hours": "24/7",
  "emergency": null
}

id is a stable {wilaya_code}-{seq} key synthesized by GeoAlgeria, unique within this dataset; the matched OSM element is kept as refs.osm. name is the best available display name and is null for unnamed points, which the app titles by type instead. type carries bilingual labels. speciality, address, phone and opening_hours come straight from OSM (null when the tags are absent). sector is "public"/"private" only when signalled, else null. geo_precision is "exact" for a surveyed OSM node or "approximate" for a building/area centroid (1,059 and 821 respectively), geo_method records which. wilaya_code joins to GeoAlgeria's wilaya_code.

Commune/wilaya linkage is derived, not from the source. OpenStreetMap does not carry Algerian administrative codes. GeoAlgeria attaches wilaya_code by point-in-polygon against the 69 wilaya boundaries, then commune_code and commune as the nearest centroid within that wilaya, so the join can never cross a wilaya boundary. Wilaya is effectively exact; commune is best-effort (centroid proximity, not polygon containment).

Need the administrative divisions too?

For wilayas, dairas, and communes, use the main geoalgeria package, it's how you turn a facility's commune_code into a polygon or centroid. Use @geoalgeria/cliniques when you only need the care facilities.

Source & method

Run npm run fetch to regenerate every output. It:

1. queries OpenStreetMap (Overpass) for amenity/healthcare=clinic, healthcare=centre and amenity/healthcare=hospital inside Algeria;
2. classifies each record from the French and Arabic name, excluding the registry tier, cabinets and paramedical schools (see the table above);
3. de-duplicates the same facility mapped as both a node and a building;
4. attaches wilaya by polygon containment and commune by nearest centroid inside it.

The raw source pull is captured under sources/cliniques/, so a rebuild never depends on Overpass being up.

License & attribution

Package code is MIT. The data is from OpenStreetMap: © OpenStreetMap contributors, licensed under the ODbL 1.0. If you use or redistribute this dataset, you must attribute OpenStreetMap contributors and keep derived databases under a compatible license.

Verify against official sources for authoritative information. This dataset is provided for reference and to power GeoAlgeria.

API docs & field reference → · Browse all packages →

Made by Yasser's Studio · LinkedIn · X · [email protected]