Commons2Data

Transpose the semantics of Wikimedia Commons categories into Wikidata statements

The end-goal is to have all the knowledge stored as categories and the category tree synced with Wikidata. Imagine a world in which we would have dynamic, request-based, multilingual categories instead of static, English-but-not-always, "1880s paintings in France of women in red dresses facing right" categories.

How it works

From a Commons category, we

• generate the list of items "within" this category, through interwiki and {{artwork}} parameters;
• translate the semantics of the category into facts (by a human);
• generate the missing statements;
• add that to Wikidata.

Tools used

• pywikibot, commonsedge

What it needs

• Moving to Tool Labs and working with dumps
• Improving the algorithm (mixing reading and writing ? Multithread ? Better tree exploration ?)
• Creating a UI to collect the category->facts translation
• Dynamic facts within the tree (a daughter category has all the statements of its mother + some)