It uses 1:1000 random sampling, so I have to count the
log events and
multiply by 1000 to get a good estimation. Am I missing something?
Quite a bit actually. Mostly that reporting is only available to "some"
browsers (the majority but not all) but also only the main document is
counted and a pageview is more than the request of the main document. For
example, you will not get all 301s/302s or images and there are many, many
other details.
See pageview definition:
https://meta.wikimedia.org/wiki/Research:Page_view
The good source for Recent pageview data is hadoop, going back a bit the
well-loved webstatscollector files provide that info:
http://dumps.wikimedia.org/other/pagecounts-all-sites/
On Wed, Jan 7, 2015 at 11:12 PM, Gergo Tisza <gtisza(a)wikimedia.org> wrote:
On Wed, Jan 7, 2015 at 5:59 PM, Nuria Ruiz
<nuria(a)wikimedia.org> wrote:
Back when
MediaViewer was launched, I added a namespace parameter to
NavigationTiming to be
able to track per-namespace pageviews,
Navigation timing is heavily sampled so I am not sure you could estimate
pageviews with the scarce dataset it provides, I would say it is not
possible.
It uses 1:1000 random sampling, so I have to count the log events and
multiply by 1000 to get a good estimation. Am I missing something?
_______________________________________________
Analytics mailing list
Analytics(a)lists.wikimedia.org
https://lists.wikimedia.org/mailman/listinfo/analytics