Research:Characterization of articles by mobile readership

Created

21:42, 10 December 2015 (UTC)

Contact

Oliver Keyes

Wikimedia Foundation

Collaborators

Daniel Jarratt

no affiliation

Thomas Christie

University of Minnesota

Mikhail Popov

no affiliation

Duration:  2015-December – ??

mobile, pageview

 Open source
via GitHub

  Open data
via GitHub

---

Research:Projects

This page documents a research project in progress.
Information may be incomplete and change as the project progresses.
Please contact the project lead before formally citing or reusing results from this page.

Do people read different Wikimedia articles on mobile devices than on non-mobile devices? We investigate the identity of articles with high mobile or non-mobile pageview proportions, and explore possible explanatory features.

Methods

• Pageview API
• Wikidata for article features

Questions

• Mobile vs. non-mobile
• Function of time / article category / language?
• Time series: we know that readership drops on the weekends, and we see that the mobile proportion of all views grows on the weekend. Is that because mobile readership grows on the weekend, or does it just drop at a slower rate?

Results

Interactivity

We have APIs for Wikipedia page views as well as Wikidata properties. How about a Top Instance/Subclass-Of for the day's top articles, grouped by target entity? E.g., 500k views for all articles about humans, 480k for television shows. And this over time, in an interactive browseable format. If it highlights missing information on Wikidata, it will focus that work based on public information need, since it's derived from pageview counts.

That does sound pretty cool; difficult to make in a robust way, I suspect? Ironholds (talk) 04:27, 25 February 2016 (UTC)