- Get all the streets in Brussels from OpenStreetMap
- Use the
name:etymology:wikidatatag to query the information about the person mentionned in the streetname - Use the data from Wikidata to determine the gender (and more) of the person
The process defined here above assume that we have all the streets of Brussels in OpenStreetMap and that all the streets of Brussels in OpenStreetMap have a name:etymology:wikidata tag if the streetname mentions a person.
To make sure we have all the streets of Brussels in OpenStreetMap, I compared the streets listing from OpenStreetMap to the streets listing from UrbIS (official data from the Brussels Region).
A few streets were missing and manually added to OpenStreetMap.
At the beginning of the project (February 2019), around 5% of the streets of Brussels in OpenStreetMap had a name:etymology:wikidata tag.
Of course, not all the streets refer to a person (or an entity) and thus need a name:etymology:wikidata tag.
To link all the streets that refer to a person to the equivalent Wikidata item, we organized an event were we asked 100 people to manually find the Wikidata item (or the Wikipedia page) about the person from the streetname.
To avoid any issue and to simplify the workflow, the 100 people didn't edit Wikipedia, Wikidata, or OpenStreetMap.
The result of their work has been tagged manually in OpenStreetMap by OpenStreetMap Belgium volunteers. (work in progress).
If there is a name:etymology:wikidata tag in OpenStreetMap, the process follow that identifier to query that item from Wikidata.
We consider to be a person, the Wikidata items that are "instance of" (property P31):
In the Wikidata object, the sex or gender is defined by the property P21.
If there is no name:etymology:wikidata tag in OpenStreetMap, we can't get the gender from Wikidata.
In this case, the gender is queried from the work done during the event of the 17th February 2020. That work is available as a CSV file. Caution, that CSV file is not to be used for statistics (not cleaned)!