On Tue, Apr 14, 2020 at 4:39 PM Thad Guidry <thadguidry(a)gmail.com> wrote:
So the Venn diagramming artwork that Freebase Sets logo used is appropriate.
To explain it for you in Wikidata terms...
As a user, Freebase Sets could be used to input let's say 2 items (max of 5).
It would run queries to figure out for the user what other topics had the same
overlapping statement values (Venn union). Figure out whats common, give me other things
that share that commonality.
Ex. I enter the following items using Wikidata as example, and on each selecting the Q
id (just as Wikidata Search dropdown suggests, it worked the same way)
"Eiffel Tower"
"Bavarian Forest National Park"
I then click on "Find Set" and it would run queries on those 2 items to find
other items that shared similar statements AND their concrete values matching exactly for
all those statements (as best it could with a few simple algorithms, hints, exclusions)
Statements and Concrete values common to those 2 items (those in () parentheses might be
excluded as not useful) :
instance of
inception
named after
country
located in the administrative territorial entity
topic's main category
(official website)
(image)
Incidentally, the above 2 items would result in probably 0 results since their values
don't match on any of those statements. But it was just a quick example to show you
how you could put in very disconnected items COULD be put into a query to find SOME
overlapping relationships discovered. That was the power of Freebase Sets. It offered a
different way to explore the graph of relations.
In Wikidata terms, imagine being able to hold down CTRL and being able to interactively
"queue up" clicking all the ellipses (triple dot next) on ALL the same
statements between the 2 items, that finally runs a big combined query to find all the
matching items that share that "overlapping statement set". (Venn union)
But it could do this with more than just 2 items to give you some really cool and
interesting result overlaps you normally would not discover or know about!
Technically, it's actually not that hard to put a Lab tool together that would be
able to mimic Freebase Sets. Some of our existing Example Queries in WQS are basically
manually curated Freebase Sets. But a tool could automate the discovery of overlapping
statements between X items.
Thanks for elaborating! That sounds interesting and worth some
exploring I'd say if someone has a bit of spare time.
Cheers
Lydia
--
Lydia Pintscher -
http://about.me/lydia.pintscher
Product Manager for Wikidata
Wikimedia Deutschland e.V.
Tempelhofer Ufer 23-24
10963 Berlin
www.wikimedia.de
Wikimedia Deutschland - Gesellschaft zur Förderung Freien Wissens e. V.
Eingetragen im Vereinsregister des Amtsgerichts Berlin-Charlottenburg
unter der Nummer 23855 Nz. Als gemeinnützig anerkannt durch das
Finanzamt für Körperschaften I Berlin, Steuernummer 27/029/42207.