Wikimedia CEE Meeting 2017/Programme/Submission/The rise and fall of bot-created articles in Swedish Wikipedia, a short summary with focus on articles about geographic places in Central and Eastern Europe

The rise and fall of bot-created articles in Swedish Wikipedia, a short summary with focus on articles about geographic places in Central and Eastern Europe

edit

Name(s) and/or username(s)

edit
Lars Aronsson User:LA2
Wikimedia Sverige

Topic

edit
  • Regional/international wiki projects (New CEE wide projects/contest)
  • Volunteer support and Volunteer recruitment
  • (?) Content Contests (Article writing contests/challenges, Photo contests)


Type of submission (Please choose one)

edit
  • Lightning talk

Summary

edit

Why is the Swedish Wikipedia the second largest and what can we do about it?

/draft/

Swedish Wikipedia's robot-created articles about geographic places

The creation of robot-created articles has been controversial from the start. Using good biological databases as a source, it was possible to create articles of acceptable quality about spiders and fungi. Articles about geographic places that were based on the Geonames database were not of acceptable quality, however, and the creation of such articles stopped at country code MX (Mexico). Articles where created about places in these countries in Central and Eastern Europe. Numbers show the number of articles still remaining in each category at the end of June 2017.

  2018 Åland      cat 
  2673 Albania    cat 
  1666 Armenia    cat 
 15964 Austria    cat	
  3441 Azerbaijan cat 
  2559 Belarus    cat 
 22284 Bosnia and Hercegovina  cat 
  4755 Bulgaria   cat 
  8025 Czech Republic  cat 
  1784 Cyprus     cat 
  1640 Estonia    cat 
 52502 Finland    cat	
  2679 Georgia    cat 
 12072 Greece     cat 
  4470 Hungary    cat 
  1859 Kosovo     cat 
   559 Latvia     cat 
   835 Lithuania  cat 
   244 Moldova    cat 
  6482 Montenegro cat 

A typical article of low quality is Apita, some kind of stream (river, creek) in Belarus. The article says it is a stream in western Belarus, but not how large or how important. Geonames is the only source reference given. However, this particular river does have articles on both Belarusian and Russian Wikipedia, so it should be possible to get more information from there.

Another article Boloto Nadel describes a marsh/swamp in southern Belarus, but has no articles in other languages of Wikipedia (except for Cebuano, which was created by the same bot), and provides no information about size or significance. Such an article probably needs to be deleted.

Both articles were created in December 2015, but the bot still continued to create articles for other countries, including Cyprus in February 2016, Greece in May 2016, Moldova in October 2016, and Mexico in November 2016.

When the bot was ordered to stop by community consensus, some drama played out and the bot operator left Wikipedia.

One contributor writes about cleaning up the articles about places in Estonia: For close to a year, I have gradually processed articles about 5-7 of Estonia's 15 provinces, possibly more than a thousand articles. Most of them describe topics that are relevant. Some duplicate articles and unverifiable articles (names not mentioned on Estonian survey maps) have been deleted. Many articles needed to be renamed because the title contained English words (parish, port, river, etc.).

Expected outcomes

edit

International understanding of this Swedish dilemma.

Duration (without Q&A)

edit

5 minutes

Specific requirements

edit

Slides or further information

edit

Interested attendees (Please add yourself, and you may indicate your questions to the presenter).

edit
  1. Gikü (talk) 22:49, 20 August 2017 (UTC)[reply]
  2. --Oop (talk) 16:29, 20 September 2017 (UTC)[reply]