Since 2008 Wikimedia collects pageview counts for most pages on nearly all wikis. A longstanding request of stakeholders (editors, researchers, GLAM advocates) has been to publish similar counts for media files: images, sounds, videos. A major obstacle to effectuate this was the existing traffic data collecting software. Webstatscollector simply couldn't be scaled up further without incurring huge costs. In 2014 WMF engineers rolled out a new Hadoop based infrastructure, which makes it possible to collect raw request counts for media files. So a few months after releasing extended pageview counts (with mobile/zero added), the time has come to produce similar data dumps for media files.
https://www.mediawiki.org/wiki/Requests_for_comment/Media_file_request_count s
Please comment onwiki.
Thanks
Erik Zachte
Hoi, Pageviews are associated with articles. Articles are associated with items. Have experiments been done to learn the page views per item? Thanks, GerardM
On 26 November 2014 at 19:06, Erik Zachte ezachte@wikimedia.org wrote:
Since 2008 Wikimedia collects pageview counts for most pages on nearly all wikis. A longstanding request of stakeholders (editors, researchers, GLAM advocates) has been to publish similar counts for media files: images, sounds, videos. A major obstacle to effectuate this was the existing traffic data collecting software. Webstatscollector simply couldn't be scaled up further without incurring huge costs. In 2014 WMF engineers rolled out a new Hadoop based infrastructure, which makes it possible to collect raw request counts for media files. So a few months after releasing extended pageview counts (with mobile/zero added), the time has come to produce similar data dumps for media files.
https://www.mediawiki.org/wiki/Requests_for_comment/Media_file_request_count s
Please comment onwiki.
Thanks
Erik Zachte
Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
wikitech-l@lists.wikimedia.org