Thanks to the work of several contributors, the usage of name:fur tag in OpenStreetMap is constantly increasing and there is now a pretty good coverage of Friuli region at place level (and even down to street-level in some parts of it). This tag represents a name in Friulian language, mainly spoken in north-east Italy.

This tag is visible today in a dedicated rendering (which I created) on mapefurlane.eu and in the various multi-lingual maps like the great Americana.

Apart from the maps, another interesting use of this data is to have lists of toponyms. For Friulian language, there is already a list provided by an official organization, but besides not being very detailed, it’s also quite static (PDF). So I wanted to have something better than that by leveraging all OSM data.

My main goal then was to create a searchable web page that could show the toponyms of Friuli historical region, not only in Friulian, but also in all recognized languages, i.e. Friulian, Italian, German and Slovenian. The list should be updated automatically from OSM data.

What is DuckDB

DuckDB is an open source in-process SQL OLAP database management system. It has a builtin spatial extension that can import PBF files which can be then queried using SQL syntax.

Using DuckDB with OpenStreetmap

Since I already had a daily pipeline (GitHub Action) to update Mape Furlane that used a PBF extract as input, integrating DuckDB there was immediate: just adding the single binary, loading the PBF and then all the data was available for querying.

From that I added a few SQL queries visible on https://github.com/andreadecorte/mapefurlane/tree/main/scripts/duckdb that generate 2 output files: 1. all places with name:fur tag and all the other name: tags (name:it, name:de, name:sl) 2. all places in the area without a name:fur tag with a Markdown output (to help increase the coverage)

The output of the first query is stored in a CSV file which is then displayed in a sortable and searchable table visible here. This is as far as I know the best multi-lingual list available online for toponyms in Friuli region. And as mentioned, it has daily updates from OSM data. As of August 2026, it shows 2370 place names.

Conclusion

This is just scratching the surface with a simple example, but DuckDB is really an awesome tool that can be used to analyze OSM data without complicated setups. Everything described in this entry is available in https://github.com/andreadecorte/mapefurlane if you are interested in digging deeper.