So, first off: Ironholds made all the numbers used in this metrics meeting available in the tool at http://pentaho.wmflabs.org/pentaho/Home
I'm not going to repost the username/password here, but find me (or Ironholds?) on IRC if you're interested in exploring the data.
http://pentaho.wmflabs.org/pentaho/Home <username>,<password> - > create new -> new Saiku Analytics -> v 0.3
Ok. With that said, here are some thoughts about the numbers, with some copy & paste from IRC:
Q: RoanKattouw: Ironholds: Re the India language graph (97% of hits from India being to enwiki), we are now idly wondering what places are more diverse in those terms RoanKattouw: Like, maybe the USA? RoanKattouw: Is the Spanish- speaking internet more than 3% of the US internet? RoanKattouw: cscott: Basically my question is, what is the % of enwiki hits in the US. Apparently for India it's 97%
A: zhwiki and/or eswiki are the top non-enwiki sites in the US; they account for about 1% of traffic. See https://en.wikipedia.org/wiki/User:Cscott/2014_December_metrics and cscott-us-proj.saiku on pentaho.
Q: cscott: also i'm very curious about, say, the rise of iran traffic -- is that to enwiki or fawiki? cscott: in general, is the global south reading enwiki? or is mobile traffic to the local wikis exploding?
A: Almost all due to enwiki traffic. See https://en.wikipedia.org/wiki/User:Cscott/2014_December_metrics cscott-ir-proj-2.saiku on pentaho, and attached graph.
Q: cscott-free: another question: does the decrease in latin america correspond to a decrease in the eswiki project?
A: The presented slide listed the following among the "Top Decliners":
Country, Month page views (billion), Annual growth rate Ecuador, 0.04, -30.5% Venezuela, 0.09, -28.0% Portugal, 0.05, -23.8% Mexico, 0.34, -23.2% Colombia, 0.14, -23.2% Chile, 0.08, -22.3% Brazil, 0.32, -21.0% Peru, 0.06, -17.8%
Countries in the top 25% by total human PVs as of October 2014; annual growth rates based on linear model (May 2013-October 2014)
I'm still working on figuring out the answer to this one. As far as I can tell, eswiki page views are pretty flat, and eswiki page views in Ecuador (for instance) are down a little, but now by 30% annually. So there's something mysterious here.
Possibly related: commons page views in latin america dropped sharply starting in 2014-06, after mediaviewer was turned on. But that doesn't seem to be quite enough. --scott
This is very important: pentaho.wmflabs.org was a hackathon project and it's running a community edition server out of a jar under my username. This is about as far from "ready to share" or "production" as you can get. We all love that it provides valuable information, so our team is working on productionizing both the Pentaho server and the pipeline that shapes the data. But again - *this is just a proof of concept, it most definitely will go away*.
On Thu, Dec 4, 2014 at 6:00 PM, C. Scott Ananian cananian@wikimedia.org wrote:
So, first off: Ironholds made all the numbers used in this metrics meeting available in the tool at http://pentaho.wmflabs.org/pentaho/Home
I'm not going to repost the username/password here, but find me (or Ironholds?) on IRC if you're interested in exploring the data.
http://pentaho.wmflabs.org/pentaho/Home <username>,<password> - > create new -> new Saiku Analytics -> v 0.3
Ok. With that said, here are some thoughts about the numbers, with some copy & paste from IRC:
Q: RoanKattouw: Ironholds: Re the India language graph (97% of hits from India being to enwiki), we are now idly wondering what places are more diverse in those terms RoanKattouw: Like, maybe the USA? RoanKattouw: Is the Spanish- speaking internet more than 3% of the US internet? RoanKattouw: cscott: Basically my question is, what is the % of enwiki hits in the US. Apparently for India it's 97%
A: zhwiki and/or eswiki are the top non-enwiki sites in the US; they account for about 1% of traffic. See https://en.wikipedia.org/wiki/User:Cscott/2014_December_metrics and cscott-us-proj.saiku on pentaho.
Q: cscott: also i'm very curious about, say, the rise of iran traffic -- is that to enwiki or fawiki? cscott: in general, is the global south reading enwiki? or is mobile traffic to the local wikis exploding?
A: Almost all due to enwiki traffic. See https://en.wikipedia.org/wiki/User:Cscott/2014_December_metrics cscott-ir-proj-2.saiku on pentaho, and attached graph.
Q: cscott-free: another question: does the decrease in latin america correspond to a decrease in the eswiki project?
A: The presented slide listed the following among the "Top Decliners":
Country, Month page views (billion), Annual growth rate Ecuador, 0.04, -30.5% Venezuela, 0.09, -28.0% Portugal, 0.05, -23.8% Mexico, 0.34, -23.2% Colombia, 0.14, -23.2% Chile, 0.08, -22.3% Brazil, 0.32, -21.0% Peru, 0.06, -17.8%
Countries in the top 25% by total human PVs as of October 2014; annual growth rates based on linear model (May 2013-October 2014)
I'm still working on figuring out the answer to this one. As far as I can tell, eswiki page views are pretty flat, and eswiki page views in Ecuador (for instance) are down a little, but now by 30% annually. So there's something mysterious here.
Possibly related: commons page views in latin america dropped sharply starting in 2014-06, after mediaviewer was turned on. But that doesn't seem to be quite enough. --scott
-- (http://cscott.net)
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics
C. Scott Ananian, 05/12/2014 00:00:
RoanKattouw: Ironholds: Re the India language graph (97% of hits from India being to enwiki),
97 %? I remembered much less. https://stats.wikimedia.org/wikimedia/squids/SquidReportPageViewsPerCountryB... Ok, I forgot over 20 % goes to www.wikipedia.org, which is totally crazy. We should maybe redirect that to http://wikimedia.in/wikipedia.html in India ;-) until we figure out something for language selection.
we are now idly wondering what places are more diverse in those terms
Pretty much any country in the world except USA and UK, according to the breakdown above. :P All the others have 5 % or more in "Other" and non-first languages.
Nemo
I've updated https://en.wikipedia.org/wiki/User:Cscott/2014_December_metrics with some more graphs and analysis.
I am unable to reproduce the decline rates reported in the metrics meeting. --scott
On 4 December 2014 at 23:00, C. Scott Ananian cananian@wikimedia.org wrote:
Possibly related: commons page views in latin america dropped sharply starting in 2014-06, after mediaviewer was turned on. But that doesn't seem to be quite enough.
Idle thought...
It used to be that, when an image was clicked, it would generate a pageview of a local File: page (for Commons images, usually a "dummy" local page). On some projects, it would instead go to the corresponding Commons file page.
However, the use of mediaviewer means that a significant fraction of clicks on images will not lead to a new pageview; the user is satisfied with the lightbox, closes it in place, and stays on the page.
In the eswiki case, this seems to be a very convincing explanation for the Commons drop. As a project which has not allowed local uploads for a long time, I believe any image links went straight to Commons rather than to the local File page used on enwiki.
Which leads to an obvious question - on other projects, such as enwiki or frwiki, have we accounted for a drop in *file page* views? What do overall pageview numbers look like using, say, just mainspace/ns0?
Idle thought...
It used to be that, when an image was clicked, it would generate a pageview of a local File: page (for Commons images, usually a "dummy" local page). On some projects, it would instead go to the corresponding Commons file page.
However, the use of mediaviewer means that a significant fraction of clicks on images will not lead to a new pageview; the user is satisfied with the lightbox, closes it in place, and stays on the page.
In the eswiki case, this seems to be a very convincing explanation for the Commons drop. As a project which has not allowed local uploads for a long time, I believe any image links went straight to Commons rather than to the local File page used on enwiki.
Which leads to an obvious question - on other projects, such as enwiki or frwiki, have we accounted for a drop in *file page* views? What do overall pageview numbers look like using, say, just mainspace/ns0?
I like the explanation. More work would probably have to be done to prove it, but it sounds good. And it illustrates the point that fewer pageviews is not always a bad thing!
I did look at this. You can clearly see the drop in eswiki vs ptwiki if you just look at page views of commons; there is a sharp drop in 2014-06 after Media Viewer is turned on.
For the "top declines" countries I looked at (well, Mexico and Ecuador), the commons page views aren't a significant fraction of overall page views, so this doesn't change the overall stats much. It's possible that for smaller projects this might significantly affect the total, but it would have to be a project with a big ratio of image views to article reads. --scott
On Fri, Dec 5, 2014 at 3:13 PM, Dan Andreescu dandreescu@wikimedia.org wrote:
Idle thought...
It used to be that, when an image was clicked, it would generate a pageview of a local File: page (for Commons images, usually a "dummy" local page). On some projects, it would instead go to the corresponding Commons file page.
However, the use of mediaviewer means that a significant fraction of clicks on images will not lead to a new pageview; the user is satisfied with the lightbox, closes it in place, and stays on the page.
In the eswiki case, this seems to be a very convincing explanation for the Commons drop. As a project which has not allowed local uploads for a long time, I believe any image links went straight to Commons rather than to the local File page used on enwiki.
Which leads to an obvious question - on other projects, such as enwiki or frwiki, have we accounted for a drop in *file page* views? What do overall pageview numbers look like using, say, just mainspace/ns0?
I like the explanation. More work would probably have to be done to prove it, but it sounds good. And it illustrates the point that fewer pageviews is not always a bad thing!
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics
On 5 December 2014 at 20:13, Dan Andreescu dandreescu@wikimedia.org wrote:
Which leads to an obvious question - on other projects, such as enwiki or frwiki, have we accounted for a drop in *file page* views? What do overall pageview numbers look like using, say, just mainspace/ns0?
I like the explanation. More work would probably have to be done to prove it, but it sounds good. And it illustrates the point that fewer pageviews is not always a bad thing!
Hmm. A quick and dirty test - look at a single high-profile image in a consistent location and see what clickthroughs are like.
Yesterday's enwiki front-page image got 2785 pageviews.
http://stats.grok.se/en/latest/File:Guyou_doubly_periodic_projection_SW.JPG
For the same image on 4 Dec 2012, 7700.
http://stats.grok.se/en/201312/File:Point_Reyes_Lighthouse_(April_2012).jpg
And the one for 4 Dec 2012, 5500.
http://stats.grok.se/en/201212/File:Neophema_chrysostoma_mortimer_2.jpg
Mainpage views on those dates were 9.3m (2012), 12.6m (2013), 12.2m (2014); so 0.02% in 2014, 0.06% in 2013, 0.06% in 2012. Approximately one third as many front-page viewers click through to POTD in 2014 as did so in 2012/13.
Views for those images *on Commons*, meanwhile, go from 15 (2012) and 35 (2013) to 235 (2014). So further confirmation for the traffic-moves-to-Commons hypothesis (and, interestingly, confirmation that a lot fewer people click through to the description page from MediaViewer...)
This suggests we're on to something here.
Anyone able to whip up total views by namespace for enwiki versus say eswiki, month by month, over the past eighteen months?
Andrew Gray, 05/12/2014 21:08:
Which leads to an obvious question - on other projects, such as enwiki or frwiki, have we accounted for a drop in*file page* views? What do overall pageview numbers look like using, say, just mainspace/ns0?
Note, however, that file [description] requests didn't simply vanish; some are being counted as article/page requests instead. https://lists.wikimedia.org/pipermail/multimedia/2014-November/000948.html
Nemo
On 5 December 2014 at 22:39, Federico Leva (Nemo) nemowiki@gmail.com wrote:
Andrew Gray, 05/12/2014 21:08:
Which leads to an obvious question - on other projects, such as enwiki or frwiki, have we accounted for a drop in*file page* views? What do overall pageview numbers look like using, say, just mainspace/ns0?
Note, however, that file [description] requests didn't simply vanish; some are being counted as article/page requests instead. https://lists.wikimedia.org/pipermail/multimedia/2014-November/000948.html
This is certainly true - if I follow the link in that email it will be a mainspace hit and counted as such, not a file-page hit.
But if I'm viewing the original page and click the image, the page doesn't reload and (presumably) thus it won't be counted as a new hit on the underlying mainspace page. (I think. Can someone confirm this one?)
So we're still going to get a steady "drain" of overall hits, even if some are still being counted in a different way.
Andrew.
