Page MenuHomePhabricator

Analyze Demographics Surveys
Closed, ResolvedPublic

Description

Run analyses of demographic survey data.

Language / Countries with greater than 500 responses

For each language, we also compute numbers for worldwide readership.

  • Arabic: Saudi Arabia, Egypt, Iraq
  • German: Germany
  • English (Worldwide): United States, India
  • English (Africa): South Africa, Nigeria, Kenya, Egypt
  • Spanish: Spain, Mexico, Argentina, Colombia, Peru, Chile
  • Persian: Iran
  • French (Worldwide): France
  • French (Africa): Morocco, Algeria
  • Hebrew: Israel
  • Hungarian: Hungary
  • Norwegian: Norway
  • Romanian: Romania
  • Russian: Russia, Ukraine
  • Ukrainian: Ukraine
  • Chinese: Taiwan

Reader Behavior Features

  • Article: see T228319
  • Session:
    • # articles read, total time, time per article
    • referer class (internal, external search engine, external)
    • number of Wikipedia language editions visited
    • viewed Main Page (Q5296)
    • where in session survey was taken
    • desktop vs. mobile
    • loggedIn or not
  • Request:
    • localized day of week, localized time of day (convert UTC time based on timezone from IP)
    • country, continent (based on IP geolocation)

Analyses

  • Unweighted (raw) proportions for each question
  • Weighted (debiased) proportions for each question
  • Cross-tabulation -- e.g., age vs. gender
  • Content vs. Reader Gaps
    • age of respondent vs. age of biographies read
    • gender of respondent vs. gender of biographies read

Event Timeline

Isaac triaged this task as Medium priority.Jul 17 2019, 8:16 PM

I will focus on the specific analyses after processing of data, debiasing, feature generation is completed

I presented preliminary results at Wikimania: https://wikimania.wikimedia.org/wiki/2019:Research/Characterizing_Reader_Behavior_on_Wikipedia

We will hopefully release fuller results in the next several weeks after doing some additional checks / review.

Analysis complete -- at this point the focus will shift to writing up and sharing the results (which has already been ongoing for several weeks).

Intermediate results in the process of being published here: https://meta.wikimedia.org/wiki/Research:Characterizing_Wikipedia_Reader_Behaviour/Demographics_and_Wikipedia_use_cases#June_2019_Results

Isaac updated the task description. (Show Details)