SIRENE database of establishments (SIRET) - geolocated with the National Address Base (BAN)

This dataset is an enrichment of the INSEE dataset (SIRENE database of enterprises and their establishments (SIREN, SIRET)) (https://www.data.gouv.fr/en/datasets/base-sirene-des-entreprises-et-de-leurs-etablissements-siren-siret/).

This enriches the original base as follows:
- Breakdown of the StockEstablishment file by geographical grid: departments and municipalities.

• Addition of a number of columns relating to the geolocation of establishments based on the most relevant proximity score between the address indicated in the SIRENE database and the [National Address Database] (https://www.data.gouv.fr/datasets/national address database/) or the [Points of Interest of Openstreetmap] (https://www.data.gouv.fr/en/datasets/points-dinterets-openstreetmap/).
• Longitude field: longitude of the establishment
• Field "latitude": Latitude of establishment
• Field geo_score: Trust score returned by the addok geocoder (between 0 and 1, the higher the score, the more relevant the geocoding seems)
• Field geo_type: type of address found
• Field geo_address: wording of the address found
• Field geo_id: identifier of this address in the source database where it was found (BAN or POI)
• Field geo_line: which address line of the SIRENE database could be geocoded (G=geographical, D=declared, N=normalized)
• Field geo_l4: line 4 to standard AFNOR address
• Field geo_l5: line 5 to standard AFNOR address

The processing allowing the production of this dataset is carried out by Etalab. It is largely inspired by the previous work of Christian Quest available [here] (https://github.com/cquest/geocoding-spd/). This processing is based on the geocoder Addok.

This dataset is used in the search engine of [the business directory] (https://annuaire-entreprises.data.gouv.fr/) and in its API (https://api.gouv.fr/les-api/api-recherche-entreprises).