[SPARQL-rc] more detailed statistics now possible?

[Sandra Fauconnier]

Hi Magnus! I'd be curious to know if the recent changes data used to feed SPARQL-rc is usable to be split up according to certain criteria, e.g.

• percentage of bot vs human edits
• percentage of edits split down by 'type' e.g. edits to identifiers, links to other Wikimedia projects, references
• gradual increase of number of statements per item; etc…

and generally to feed into statistics over time à la BaGLAMa?

I just revisited some of the documentation we wrote after the Flemish museums' import project, specifically about statistics requests that could not be fulfilled at that time yet - but maybe now they can.

Here’s what we wrote about desired statistics on wiki.

I assume (1) and (2) (partially) now may become possible ?? Or not?

If it does become possible, I'd be interested to ask a few other Wikidata statistics enthusiasts to brainstorm together more specifically on what kind of statistics we'd love to have.

[Magnus Manske]

Hi Sandra, some basic stats are certainly possible:

• bot/human is easy
• "type" is possible, but requires some fiddling (this is encoded as text in the revision edit summary, somehow...)
• that would include increase/decrease statements etc. within the diff period
• over time statistics are of course possible, but require a database, an interface, user management, OAuth login, access as API rather than HTML pages, and so on; in other words, careful planning!
• I actually wrote a tool for Liam in that regard a long time ago, but I forgot where it is (don't laugh!), but it wasn't complete, and has most like fallen into disrepair by now

A "statistics wishlist" would be very nice for planning; the Flemish "Statistics" list is certainly a start. SPARQL-rc is a rather ad-hoc tool, I might actually write a separate one for statistics collection over time. Any name suggestions? ;-)

[Magnus Manske]

Hah, found the tool I mentioned! And you were a customer :-)

[Magnus Manske]

I think I'll just keep that tool name and modify it. Keeping my wits about me, so to speak ;-)

[Magnus Manske]

A first attempt