Automated edits/bxl-forever

Import new databrussels id for address points in Brussels

Proposal	Add the ref:databrussels=* tag to every address node in Brussels.
Rationale	Addresses in Brussels have been historically tagged with ref:UrbIS=*, which uniquely identifies the building officially hosting an address. The government no longer maintains this system and uses a new numbering scheme, part of the BeSt address register, used consistently across Belgium. Each addressable point (i.e. a house, or each mailbox inside large apartment buildings) has a unique iD. The Brussels government maintains a public website where the value of this tag can be used to retrieve any address in Brussels.
Frequency and timing	One single import now (Jan-Feb 2026), possibly in a limited number of changesets. The government publishes updates all though the year (14 updates in 2025); minor imports may come in the future to add bunches of new addresses.
Areas affected	Brussels-Capital Region (Belgium)
Size of import	About 55,800 addresses are currently mapped as nodes. Each existing node will get an extra ref:databrussels=* tag. Other addresses are tagged on buildings (about 165,000 buildings) and will not be modified here.
Source data	https://datastore.brussels/web/data/dataset/2cf42541-1813-11ef-8a81-00090ffe0001 is offered as a GeoPackage, which contains raw data.
Licence	The link above shows the licence conditions for the data. Whereas the data layer containing "cadastral parcels" has a restrictive licence, this data is NOT used for the import. The import solely use the "Addresses" vector layer. The document states this: "The other classes fall under the Creative Commons (CC-0) - no conditions for reuse, the data is in the public domain."
Method	(Already done) Download the latest GeoPackage version (currently 2026-01-01), import the Address vector layer in QGIS to transform the Lambert72 geometry into WGS84, export as GeoJSON, parse the GeoJSON file to discard child addresses (apartments inside apartment buildings) and only keep master addresses, import them in a local database, resulting in 226,448 unique addresses. Import OSM address data into the local database. (To do) Manual inspection of duplicate values in source data (currently 3 values ouf of 226,448). Compare each address with existing OSM data. Only if the address matches (FR and NL street name, single housenumber, latitude/longitude within 30 metres), issue a JOSM remote control query to add the tag to the existing object. List unmatched addresses separately for manual investigation (spelling issues, identical address elsewhere in the city, several addresses which were mapped on a single node in OSM…).
Talk	Will be discussed on the OSM.be Matrix chat, which has a room dedicated to imports and buildings/addresses.
Status	Proposed
Changeset(s)	-