This morning, I stumbled upon this Wikidata error.
I wanted to fix it, but I did not know how.
I wish there were an "intelligent autocorrect" feature built into Wikidata that could suggest a quick fix for situations like this. Modern IDEs already have such features.
Anyway. So, I started thinking of several ways to solve the problem.
One of my proposed solutions was to do a quick “global survey of Wikidata-wide usage” of P812 (academic major). Generate a list of all possible "legal" values of P812, then sort them by frequency.
So, I crafted this SPARQL query.
Using that SPARQL query, I could generate this list.
Satisfactory?
No.
I felt something big was missing.
That query only counted direct statements.
I also needed to count the other case.
So, I crafted another SPARQL query.
But the Wikidata Query Service failed to execute it due to an upstream request timeout.
So, I modified the SPARQL query so it could be executed in QLever.
It worked.
Now, I can simply browse that list to see what kinds of "academic major" values are considered valid to enter into Wikidata.