Page MenuHomePhabricator

Measure compared deletion ratios for articles created with and without Content translation
Closed, ResolvedPublic


Content translation allow users to create new articles. In order to better understand the quality of the content created with the tool it would be useful to compare the deletion ratios for both articles created with and without content translation.

We want to be able to obtain a set of statements like the following one:

On French Wikipedia the deletion ratio is 5% for articles created with Content Translation, and 27% for new articles created otherwise.

Wikis to measure

We want to capture the results for the following Wikipedias (based on the list of representative wikis with some large Wikipedia additions for reference):

  • English
  • German
  • Indonesian
  • Arabic
  • Catalan
  • Czech
  • French
  • Hebrew
  • Italian
  • Korean
  • Portuguese
  • Russian
  • Spanish
  • Tamil
  • Ukrainian

In addition, if possible, we want to capture the overall numbers for all Wikipedias.

Additional considerations

  • The results will be calculated for a given period of time. All measurements will use the same period of time. Depending on the cost of the queries, we can pick a one year, 6 months or 3 months as the time period.
  • Note that the "Non-CX deletion ratio" requires to exclude the pages created with CX. That is, calculating the deletion ratio for all new pages is not enough since that includes pages created with CX too.
  • The results and the queries used to obtain them will be published on wiki and linked from the CX analysics page.
  • The results can be expressed in a table like the one below:
WikiCX deletion ratioNon-CX deletion ratio
French Wikipedia5%27%
Spanish Wikipedia10%52%
All wikipedias10%35%