Commons:WMF support for Commons/December 2022 metrics

From Wikimedia Commons, the free media repository
Jump to navigation Jump to search
[edit]

Wikimedia Commons is Growing.


This plot shows the total number of items (articles for Wikipedia, items for Wikidata, files for Wikimedia Commons, and entries for Wikisource) registered on four Wikimedia projects in the past eight years. The content on Wikimedia Commons has grown by 400% since 2014, from around 20M files to more than 80M in 2022. This means that Wikimedia Commons is growing faster than Wikipedia.

Source and more data on Wikistats (link points to the Commons data).

When we focus on recent Commons active editors, we see that between 2019 and 2021, returning monthly active editors grew by 16%, while the new active editors grew by 12%. These value as much higher than Wikipedia (+8% and +6% respectively)

Source: Product Analytics editors data (internal), 2001-01 to 2022-05. An active editor is a registered, non-bot user who makes at least 5 edits to content namespaces in a month. See full definition on Meta.

Commons’ direct consumers may not be many…

What about Wikimedia Commons consumers? When we look at the total number of requests that come to our sites, Wikimedia Commons pages do not get a lot of direct views: about 100M requests/month, while English Wikipedia alone gets about 7B monthly pageviews.

Source: Aggregated from internal data across Wikimedia sites. Enwiki pageviews on Wikistats; Commons pageviews on Wikistats


… but the Impact of Commons is through its Reuse.

When we add Commons files to Wikipedia articles, their visibility increases massively. This plot shows the average hourly pageviews (excluding bots) on English Wikipedia content articles (namespace 0, data from March 2021). Pageviews are split by articles with images (in blue) and without images (in orange). What we see is that consistently, across hours of the day and day of the week, Wikipedia articles that have at least an image get 4 times more pageviews than unillustrated articles. In March 2021, illustrated articles received a total of 5.4B pageviews, while unillustrated articles got about 1.4B. This is not a causal effect: increased pageviews are not a consequence of the presence of images in the articles. But this is a signal of the large number of consumers of Wikimedia Commons files that make it into Wikipedia.

At the same time, when we enrich Wikipedia content with images, it becomes more engaging. We found that images are very engaging for our readers. About 1 out of 30 times we visit a page, we also click on an image: the image click-through rate is higher than 3%. These numbers are significantly higher than engagement numbers of other interactive parts of the page: citation clickthrough rate is around 0.3%, while links click-through is about 1%.

For more information and more data on readers’ engagement with images (including what images are more engaging for Wikipedia readers), check this research paper or the project meta page.

Readers would like more and better images

What would you improve about Wikipedia, if you could?
13-24 25-44 45+
Include more videos 43% 44% 38%
Make it more visual 41% 41% 37%
Add more detail in articles 38% 37% 33%
Make articles higher quality 37% 37% 30%
Include recommendations of articles relevant to your interests 37% 39% 36%
Make it easier to use when accessing on a mobile browser 30% 33% 26%
Have more articles 28% 29% 24%
Make articles shorter 20% 17% 12%
Make it easier to use when using the app 20% 21% 18%
Make it easier to use on PC 19% 19% 20%
Have more articles in your preferred language 19% 23% 17%
Make articles longer 18% 15% 12%
Reduce amount of detail in articles 13% 12% 9%


In a Global Market Research in 2020, we asked readers from 10 different countries who used Wikipedia in the last 3 months what improvements they would like to see on Wikipedia. We found that Readers, especially in emerging markets, want an improved visual experience and more multimedia-based content: the top 2 requests were “more visuals” and “more videos”. This suggests that, while Commons is very important for Wikipedia and vice-versa, there is still room for improvement.


The Visual Knowledge Gap on Wikipedia

One of the major issues of the Commons-Wikipedia ecosystem is that on average, 40% of articles on English Wikipedia remain unillustrated. This percentage is even higher (>50%) for large Wikipedia projects such as English or German Wikipedia. For most Wikipedias, between 40% and 60% of the articles are illustrated, but for example, only 1 in 4 articles in Vietnamese Wikipedia contain an image. Across languages, shorter articles, where non-textual knowledge can be very helpful to complement the lack of content, are unfortunately the most affected by the lack of images. More details in our recent analysis. The visual knowledge gap goes beyond the actual number of illustrated articles. Images on Wikipedia are missing captions and descriptions. For example, only 46% of images in English Wikipedia come with a caption text, and we estimate that only 10% have some form of alt-text, with 3% having an alt-caption that is appropriate for accessibility purposes. This lack of contextual information not only limits the accessibility of visual and textual content on Wikipedia, but it also affects the way in which images can be retrieved and reused across the web.