[Wikipedia-l] Edits by project and country of origin

Lars Aronsson lars at aronsson.se
Mon Sep 4 16:33:57 UTC 2006


Kelly Martin wrote:

> * NL and BE are language sluts.

This does not seem very likely.  I just don't see 26.7% Dutch 
contributors on the Danish Wikipedia.  Who are they?

Even if you did include some interwiki link robot in the 
statistics, 26.7% of all contributions (or contributors?) seems 
far too much.

That users in Belgium are split between French and Dutch is the 
nature of this country.  Belgium originates 1.8% of all 
contributions, which is also reasonable.

It's the number of users from the Netherlands that are far too 
many, 6.1% of all contributors, ranking as the 4th contributor 
country.  They should rank lower then contributors from France 
(5.6%) and closer to Belgium.  Perhaps they should be half their 
present number (3.1%).

I guess the odd statistics could be explained by some major ISP 
(UPC/Chello?) reporting their entire European network to be 
located in the Netherlands.  Did you check this?

Assuming that half of the reported number are real Dutchmen, that 
would put their contributions 66% in the Dutch Wikipedia, 25% in 
the English Wikipedia, and 9% in other places.  The other half of 
the alleged Dutchmen could be equally distributed among other 
European countries.  The reported Dutch contributions to other 
languages could be moved to each native country, thus the Danish 
Wikipedia would have 54.8 + 26.7 = 81.5% contributions from 
Denmark.  That makes more sense.

You need to fix the error in the geolocation database, or to 
exclude the IP addresses that you fail to locate correctly.  
Alternatively, you could count the problematic ISP as a virtual 
country of its own.  Just don't count them as Dutch people.


-- 
  Lars Aronsson (lars at aronsson.se)
  Aronsson Datateknik - http://aronsson.se



More information about the Wikipedia-l mailing list