On Fri, Oct 2, 2015 at 4:51 PM, Federico Leva (Nemo) <nemowiki(a)gmail.com>
Thad Guidry, 02/10/2015 21:44:
To my eyes, it shows that the Asia continent is
still generally void of
any useful machine-readable Knowledge, in either Freebase or Wikidata.
(or anywhere else) But this is already a known state of affairs and
probably will not improve until 1 Million USA students learn Mandarin. :)
Extrapolating from Freebase and Wikidata may not be reliable. For example,
Baidu is extracting structured microdata from web sites:
It also shows that Wikidata and Freebase have
different opinions on what's
the centre of Europe (or maybe one of the two has tons of statements on
Cape Town! too lazy to manually calculate labels on the axes).
I agree that labeled axes would be useful. Both graphs have a strong
vertical lines which line up with each other near the prime meridian
(Paris?), so I don't think there's a shift involved. The extra line is on
the Wikidata graph. I'm guessing it's Rome, not Cape Town. The Italians
are pretty keen Wikidatans, aren't they?
The Wikidata graph also seems to exhibit more and bigger strong horizontal
features than the Freebase graph, in particular one in ~875AD? which spans
half the globe.
If there's an intermediary form of the data that includes lat/lon instead
of just longitude, an interactive visualization overlaid on a world map
with a slider for date and selectable Freebase/Wikidata plots might be fun.