Discussion:
[Wikimetrics] Automated bulk collection of pageviews
Pine W
2015-07-25 16:19:15 UTC
Permalink
Hi Wikimetricians,

An off-list question from one of my colleagues gave me an idea.

Could Wikimetrics be expanded to show total pageviews of articles and total
views of images, for entire categories and/or lists of multiple articles
starting from variable dates? This would help GLAMs and affiliates to track
the viewership of content creation that we have supported.

Or are there tools other than Wikimetrics that already make this possible?

Thanks,
Pine
Dan Andreescu
2015-07-27 14:55:05 UTC
Permalink
Post by Pine W
Could Wikimetrics be expanded to show total pageviews of articles and
total views of images, for entire categories and/or lists of multiple
articles starting from variable dates?
Wikimetrics is a canned query tool. It could be modified, sure, but we
would need a public API that served the data you're trying to query.
Wikimetrics answers all its current queries from labsdb, which does not
have any pageview data. So far, this pageview data has only been available
via these archives that are added hourly:

media counts: http://dumps.wikimedia.org/other/mediacounts/daily/2015/
page views: http://dumps.wikimedia.org/other/pagecounts-all-sites/ (soon
to come, better more normalized data)
Post by Pine W
This would help GLAMs and affiliates to track the viewership of content
creation that we have supported.
Or are there tools other than Wikimetrics that already make this possible?
This quarter we are focusing on a public API that you can use to query this
kind of data on-demand. While Wikimetrics could use this data, and the
metric would be easy to write, it would be harder to extend the "cohort of
users" concept to "cohort of pages" or "category of pages". Maybe the API
we build will be easy enough to use that Wikimetrics won't be required.

I have become quite skeptical of Wikimetrics's value after some talks with
the folks who use it. It seems that instead of Wikimetrics a lot of people
would prefer just an API that takes a list of users and returns reports
with a set of standard metrics ran over a standard period of time with
standard parameters. I'd love to hear thoughts around this.
Dario Taraborelli
2015-07-27 15:12:12 UTC
Permalink
a lot of people would prefer just an API that takes a list of users and
returns reports with a set of standard metrics ran over a standard period
of time with standard parameters.

curious to hear more about this, Dan. Can we take 15 minutes to chat this
week?
Could Wikimetrics be expanded to show total pageviews of articles and
Post by Pine W
total views of images, for entire categories and/or lists of multiple
articles starting from variable dates?
Wikimetrics is a canned query tool. It could be modified, sure, but we
would need a public API that served the data you're trying to query.
Wikimetrics answers all its current queries from labsdb, which does not
have any pageview data. So far, this pageview data has only been available
media counts: http://dumps.wikimedia.org/other/mediacounts/daily/2015/
page views: http://dumps.wikimedia.org/other/pagecounts-all-sites/ (soon
to come, better more normalized data)
Post by Pine W
This would help GLAMs and affiliates to track the viewership of content
creation that we have supported.
Or are there tools other than Wikimetrics that already make this possible?
This quarter we are focusing on a public API that you can use to query
this kind of data on-demand. While Wikimetrics could use this data, and
the metric would be easy to write, it would be harder to extend the "cohort
of users" concept to "cohort of pages" or "category of pages". Maybe the
API we build will be easy enough to use that Wikimetrics won't be required.
I have become quite skeptical of Wikimetrics's value after some talks with
the folks who use it. It seems that instead of Wikimetrics a lot of people
would prefer just an API that takes a list of users and returns reports
with a set of standard metrics ran over a standard period of time with
standard parameters. I'd love to hear thoughts around this.
_______________________________________________
Wikimetrics mailing list
https://lists.wikimedia.org/mailman/listinfo/wikimetrics
--
*Dario Taraborelli *Head of Research, Wikimedia Foundation
wikimediafoundation.org • nitens.org • @readermeter
<http://twitter.com/readermeter>
Dan Andreescu
2015-07-27 15:45:51 UTC
Permalink
On Mon, Jul 27, 2015 at 11:12 AM, Dario Taraborelli <
Post by Dario Taraborelli
a lot of people would prefer just an API that takes a list of users and
returns reports with a set of standard metrics ran over a standard period
of time with standard parameters.
curious to hear more about this, Dan. Can we take 15 minutes to chat this
week?
Of course, Dario, anytime

James Hare
2015-07-27 15:40:59 UTC
Permalink
I have become quite skeptical of Wikimetrics's value after some talks with the folks who use it. It seems that instead of Wikimetrics a lot of people would prefer just an API that takes a list of users and returns reports with a set of standard metrics ran over a standard period of time with standard parameters. I'd love to hear thoughts around this.
Wikimetrics is very flexible and powerful. It is good if you know what you are doing. But those of us who are WMF grantees are required to report on a standard set of metrics—the flexibility is not necessarily useful for us. Further, it should be easy as possible to report these metrics. I would love to be able to dump a list of usernames and a time interval and get back the global metrics report.


—
James Hare
President, Wikimedia DC
http://wikimediadc.org
@wikimediadc
Loading...