San Vincenzo house numbers import
This page refers to the import of house numbers of the Comune di San Vincenzo in Livorno (LI), Italy. The data used is provided by the Comune di San Vincenzo.
Goals
As of April 2025 many addresses in the Comune di San Vincenzo are missing in OSM. There are only ca. 60 addresses present, while there are about ~8k numbers (assigned to property entrances) registered in the database of Comune di San Vincenzo.
The goal of the import is to add the missing addresses to the OSM database.
Legal
Data source site (as of April 2025): https://dati.toscana.it/dataset/toponimi_e_numeri_civici10
Type of license: Italian Open Data License v2.0
Licence compatibility : compatible
Dedicated upload account: https://www.openstreetmap.org/user/szydzio-imports
The import has been announced on the Italian OSM mailing list.
Dataset
The dataset is available at the Regione Toscana OpenData page, precisely at this page.
The data is presented among others in KML and CSV formats and these formats will be used for the import.
The KML data (accesso_esterno.kml) has the coordinates for each property entrance (aka accesso esterno) combined with its unique code (cvdiae). The KML data (Toponimo_Stradale.kml) has the multipolygons for every street/square combined with unique code of the street/square (tp_str_cod) and its name (tp_str_nom).
The CSV data (numero_civico.csv) has a list of numbers (civico_num) and subordinates (aka esponente) (civico_sub) assigned to every street (tp_str).
Import process
Basically 6 steps will be made for the import of the data:
- downloading the data
- exctracting coordinates of the housenumbers together with its unique code
- exctracting street/square names together with their unique code
- combining the housenumbers with coordinates and street/square names
- adding the city name,
- importing all the punctual elements of the dataset as nodes into OSM using opendata plug-in.
The first step will be made using Mozilla Firefox, next steps will be made with Notepad, Excel and JOSM.
Particularly in the fourth step, for each point:
- for each point the street name will be changed to the name of the street currently preexistenting in OSM (e.g. VIA ROMA will be changed to Via Roma) in order not to produce addresses unable to be found later by Nominatim (the house number attached to an existing street nearby should have exactly the same name as the street in addr:street)
- street names will be saved as addr:street, and the house numbers as addr:housenumber,
- in case of hamlets (Frazione/Località) the addr:place key will be used instead of addr:street, expanding possible abbreviation to full names (e.g. "Fraz. Chiave" => "Frazione Chiave")
- the letter part of the addr:housenumber (esponente) will be changed to lowercase (e.g. 30A => 30a)
In the fifth step, for each point, the keys will be added:
- addr:city with value "San Vincenzo";
After step fifth the ready *.osm file would be as follows.
In the sixth step an upload of the points as nodes (according to the IT convention) will be executed in bunches of no more than 500 elements in order to carefully cope with the data and revert them easily using JOSM's Reverter plug-in, should something go wrong.
People involved
- data preparing & import process - szydzio via import-account
Problems and improvements
I expect some problems with the import like e.g.:
- some address may already be present in the OSM database - they will be left unchanged, the data will be filtered with JOSM's validator in order not to produce duplicated numbers with this import,
- if the addresses already present in OSM are not correct (e.g. have a typo in addr:street or missing addr:street) they will be corrected,
- some house number may correspond to a roads/squares or places that are currently missing in OSM - they will be mapped separately (using normal account and/or with the help of (local) mappers, who offered their support).
QA
After the import possible errors will be checked with Nominatim QA , particularly with the Suspicious addr:street tag layer and OSM Inspector.
Schedule
The changesets will be uploaded starting by the end of April 2025.