Hello,
First of all, thank you for providing such a wealth of information regarding Wikipedia usage. This is really interesting.
I was browsing the pageviews visualization provided by wmflabs and noticed a particular case which surprised. The FR page for *Batman v Superman *seems to have received no visit whasoever before 4/29/2016:
https://tools.wmflabs.org/pageviews/?project=fr.wikipedia.org&platform=a...
However, the page existed as early as 2014:
https://fr.wikipedia.org/w/index.php?title=Batman_v_Superman_:_L%27Aube_de_l...
and similar pages in IT or ES show quite a lot of views:
https://tools.wmflabs.org/pageviews/?project=it.wikipedia.org&platform=a... https://tools.wmflabs.org/pageviews/?project=es.wikipedia.org&platform=a...
Any idea why the complete lack of pageviews for the FR page? Even if for some reason the page was not very popular in FR, shouldn't it have received at least some views?
Many thanks in advance for any ideas, Gheorghe
Hi Gheorghe,
Thanks for your message and support :) The reason the page had no views before 4/29/2016is because its name has changed that day [1], and we collect pageviews by name.
You can find the raw data using direct API calls ([2] for new name, [3] for old name, only capital J differs ...), but I have not found a way to view it properly with the pageviews tool. Being able to request titles that are not current pages could be an interesting feature request -- Ping @MusikAnimal ;)
Cheers ! Joseph
[1] https://fr.wikipedia.org/w/index.php?title=Batman_v_Superman_%3A_L%27Aube_de... [2] https://wikimedia.org/api/rest_v1/metrics/pageviews/per-article/fr.wikipedia... [3] https://wikimedia.org/api/rest_v1/metrics/pageviews/per-article/fr.wikipedia...
On Thu, Apr 13, 2017 at 4:10 PM, Gheorghe Postelnicu < gheorghe.postelnicu@gmail.com> wrote:
Hello,
First of all, thank you for providing such a wealth of information regarding Wikipedia usage. This is really interesting.
I was browsing the pageviews visualization provided by wmflabs and noticed a particular case which surprised. The FR page for *Batman v Superman *seems to have received no visit whasoever before 4/29/2016:
https://tools.wmflabs.org/pageviews/?project=fr. wikipedia.org&platform=all-access&agent=user&range=last- year&pages=Batman_v_Superman_:_L%27Aube_de_la_justice
However, the page existed as early as 2014:
https://fr.wikipedia.org/w/index.php?title=Batman_v_ Superman_:_L%27Aube_de_la_justice&dir=prev&action=history
and similar pages in IT or ES show quite a lot of views:
https://tools.wmflabs.org/pageviews/?project=it. wikipedia.org&platform=all-access&agent=user&range=last- year&pages=Batman_v_Superman:_Dawn_of_Justice https://tools.wmflabs.org/pageviews/?project=es. wikipedia.org&platform=all-access&agent=user&range=last- year&pages=Batman_v_Superman:_Dawn_of_Justice
Any idea why the complete lack of pageviews for the FR page? Even if for some reason the page was not very popular in FR, shouldn't it have received at least some views?
Many thanks in advance for any ideas, Gheorghe
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics
TL;DR, see [1] for the chart on the old page and new page.
By default the search uses the native autocompletion, which is usually what you want. E.g. if you search for "Barak Obama" you will see very few pageviews when most people actually wanted "Barack Obama". You can change this behaviour by going to the Settings > "Search method" > "Autocompletion including redirects". I still had trouble pulling it up with that setting though, something I'm not sure I can fix, as it seems to be an issue with the search API. In most cases however "Autocompletion including redirects" should reveal what you want, otherwise you could try "No autocompletion".
Anyway here's data for both the old location and new: [1]. There is some overlap because when you browse to a redirect, it will count as a pageview for the redirect itself, and not for the target article. If you want all pageviews, including redirects, consider using the "Redirect Views" tool: [2]
You probably are thinking to yourself that you always want redirects to be included, but doing so does not ensure completely accurate data. For instance, those redirects could have at some point in the past gone to a different article. So you'd then be getting (probably slightly) inaccurate numbers. There are talks of tackling this issue in the API itself, see [3]. In addition some pages have lots of redirects which could noticeably slow down the tool.
Another possible solution is to query the page move log. This is something I hope to actually do at some point, you can follow the ticket on Phabricator for updates [4]. This may be quite challenging, since there are some edge cases where the move log will steer you wrong and you end up querying the wrong pages.
Sorry for my excessive tech jargon, if it comes off that way :) Until we have a better way of handling this, first make sure you have the "Search method" set to "Autocomplete including redirects" or "No autocompletion", then simply type in the old page title.
Hope this helps,
~MA
[1] https://tools.wmflabs.org/pageviews/?project=fr.wikipedia.org&platform=a... [2] https://tools.wmflabs.org/redirectviews/?project=fr.wikipedia.org&platfo... [3] https://phabricator.wikimedia.org/T121912 [4] https://phabricator.wikimedia.org/T141332
On Thu, Apr 13, 2017 at 11:10 AM, Joseph Allemandou < jallemandou@wikimedia.org> wrote:
Hi Gheorghe,
Thanks for your message and support :) The reason the page had no views before 4/29/2016is because its name has changed that day [1], and we collect pageviews by name.
You can find the raw data using direct API calls ([2] for new name, [3] for old name, only capital J differs ...), but I have not found a way to view it properly with the pageviews tool. Being able to request titles that are not current pages could be an interesting feature request -- Ping @MusikAnimal ;)
Cheers ! Joseph
[1] https://fr.wikipedia.org/w/index.php?title=Batman_v_ Superman_%3A_L%27Aube_de_la_justice&type=revision&diff= 125734639&oldid=125727682 [2] https://wikimedia.org/api/rest_v1/metrics/pageviews/per- article/fr.wikipedia/all-access/user/Batman_v_Superman_ %3A_L'Aube_de_la_justice/daily/2015010100/2016123100 [3] https://wikimedia.org/api/rest_v1/metrics/pageviews/per- article/fr.wikipedia/all-access/user/Batman_v_Superman_ %3A_L'Aube_de_la_Justice/daily/2015010100/2016123100
On Thu, Apr 13, 2017 at 4:10 PM, Gheorghe Postelnicu < gheorghe.postelnicu@gmail.com> wrote:
Hello,
First of all, thank you for providing such a wealth of information regarding Wikipedia usage. This is really interesting.
I was browsing the pageviews visualization provided by wmflabs and noticed a particular case which surprised. The FR page for *Batman v Superman *seems to have received no visit whasoever before 4/29/2016:
https://tools.wmflabs.org/pageviews/?project=fr.wikipedia. org&platform=all-access&agent=user&range=last-year&pages= Batman_v_Superman_:_L%27Aube_de_la_justice
However, the page existed as early as 2014:
https://fr.wikipedia.org/w/index.php?title=Batman_v_Superman _:_L%27Aube_de_la_justice&dir=prev&action=history
and similar pages in IT or ES show quite a lot of views:
https://tools.wmflabs.org/pageviews/?project=it.wikipedia. org&platform=all-access&agent=user&range=last-year&pages= Batman_v_Superman:_Dawn_of_Justice https://tools.wmflabs.org/pageviews/?project=es.wikipedia. org&platform=all-access&agent=user&range=last-year&pages= Batman_v_Superman:_Dawn_of_Justice
Any idea why the complete lack of pageviews for the FR page? Even if for some reason the page was not very popular in FR, shouldn't it have received at least some views?
Many thanks in advance for any ideas, Gheorghe
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics
-- *Joseph Allemandou* Data Engineer @ Wikimedia Foundation IRC: joal
Thanks a lot MusikAnimal for those explanations, this helps (me!) a lot :) Joseph
On Thu, Apr 13, 2017 at 5:59 PM, Leon Ziemba musikanimal@wikimedia.org wrote:
TL;DR, see [1] for the chart on the old page and new page.
By default the search uses the native autocompletion, which is usually what you want. E.g. if you search for "Barak Obama" you will see very few pageviews when most people actually wanted "Barack Obama". You can change this behaviour by going to the Settings > "Search method" > "Autocompletion including redirects". I still had trouble pulling it up with that setting though, something I'm not sure I can fix, as it seems to be an issue with the search API. In most cases however "Autocompletion including redirects" should reveal what you want, otherwise you could try "No autocompletion".
Anyway here's data for both the old location and new: [1]. There is some overlap because when you browse to a redirect, it will count as a pageview for the redirect itself, and not for the target article. If you want all pageviews, including redirects, consider using the "Redirect Views" tool: [2]
You probably are thinking to yourself that you always want redirects to be included, but doing so does not ensure completely accurate data. For instance, those redirects could have at some point in the past gone to a different article. So you'd then be getting (probably slightly) inaccurate numbers. There are talks of tackling this issue in the API itself, see [3]. In addition some pages have lots of redirects which could noticeably slow down the tool.
Another possible solution is to query the page move log. This is something I hope to actually do at some point, you can follow the ticket on Phabricator for updates [4]. This may be quite challenging, since there are some edge cases where the move log will steer you wrong and you end up querying the wrong pages.
Sorry for my excessive tech jargon, if it comes off that way :) Until we have a better way of handling this, first make sure you have the "Search method" set to "Autocomplete including redirects" or "No autocompletion", then simply type in the old page title.
Hope this helps,
~MA
[1] https://tools.wmflabs.org/pageviews/?project=fr. wikipedia.org&platform=all-access&agent=user&range=last- year&pages=Batman_v_Superman_:_L%27Aube_de_la_Justice| Batman_v_Superman_:_L%27Aube_de_la_justice https://tools.wmflabs.org/pageviews/?project=fr.wikipedia.org&platform=all-access&agent=user&range=last-year&pages=Batman_v_Superman_:_L%27Aube_de_la_Justice%7CBatman_v_Superman_:_L%27Aube_de_la_justice [2] https://tools.wmflabs.org/redirectviews/?project=fr. wikipedia.org&platform=all-access&agent=user&range=last- year&sort=views&direction=1&view=list&page=Batman_v_ Superman_:_L%27Aube_de_la_justice [3] https://phabricator.wikimedia.org/T121912 [4] https://phabricator.wikimedia.org/T141332
On Thu, Apr 13, 2017 at 11:10 AM, Joseph Allemandou < jallemandou@wikimedia.org> wrote:
Hi Gheorghe,
Thanks for your message and support :) The reason the page had no views before 4/29/2016is because its name has changed that day [1], and we collect pageviews by name.
You can find the raw data using direct API calls ([2] for new name, [3] for old name, only capital J differs ...), but I have not found a way to view it properly with the pageviews tool. Being able to request titles that are not current pages could be an interesting feature request -- Ping @MusikAnimal ;)
Cheers ! Joseph
[1] https://fr.wikipedia.org/w/index.php?title=Batman_v_Superman _%3A_L%27Aube_de_la_justice&type=revision&diff=125734639&oldid=125727682 [2] https://wikimedia.org/api/rest_v1/metrics/pageviews/per- article/fr.wikipedia/all-access/user/Batman_v_Superman_%3A_ L'Aube_de_la_justice/daily/2015010100/2016123100 [3] https://wikimedia.org/api/rest_v1/metrics/pageviews/per- article/fr.wikipedia/all-access/user/Batman_v_Superman_%3A_ L'Aube_de_la_Justice/daily/2015010100/2016123100
On Thu, Apr 13, 2017 at 4:10 PM, Gheorghe Postelnicu < gheorghe.postelnicu@gmail.com> wrote:
Hello,
First of all, thank you for providing such a wealth of information regarding Wikipedia usage. This is really interesting.
I was browsing the pageviews visualization provided by wmflabs and noticed a particular case which surprised. The FR page for *Batman v Superman *seems to have received no visit whasoever before 4/29/2016:
https://tools.wmflabs.org/pageviews/?project=fr.wikipedia.or g&platform=all-access&agent=user&range=last-year&pages=Batma n_v_Superman_:_L%27Aube_de_la_justice
However, the page existed as early as 2014:
https://fr.wikipedia.org/w/index.php?title=Batman_v_Superman _:_L%27Aube_de_la_justice&dir=prev&action=history
and similar pages in IT or ES show quite a lot of views:
https://tools.wmflabs.org/pageviews/?project=it.wikipedia.or g&platform=all-access&agent=user&range=last-year&pages=Batma n_v_Superman:_Dawn_of_Justice https://tools.wmflabs.org/pageviews/?project=es.wikipedia.or g&platform=all-access&agent=user&range=last-year&pages=Batma n_v_Superman:_Dawn_of_Justice
Any idea why the complete lack of pageviews for the FR page? Even if for some reason the page was not very popular in FR, shouldn't it have received at least some views?
Many thanks in advance for any ideas, Gheorghe
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics
-- *Joseph Allemandou* Data Engineer @ Wikimedia Foundation IRC: joal
This is awesome information, many thanks Leon and Joseph.
It seems that the information you mentioned is also available in the redirects API, i.e.
https://fr.wikipedia.org/w/api.php?format=xml&action=query&titles=Ba...
which covers all but the last 2 of the links you sent, e.g.
https://tools.wmflabs.org/redirectviews/?project=fr.wikipedia.org&platfo...
Best, Gheorghe
On Thu, Apr 13, 2017 at 5:59 PM, Leon Ziemba musikanimal@wikimedia.org wrote:
TL;DR, see [1] for the chart on the old page and new page.
By default the search uses the native autocompletion, which is usually what you want. E.g. if you search for "Barak Obama" you will see very few pageviews when most people actually wanted "Barack Obama". You can change this behaviour by going to the Settings > "Search method" > "Autocompletion including redirects". I still had trouble pulling it up with that setting though, something I'm not sure I can fix, as it seems to be an issue with the search API. In most cases however "Autocompletion including redirects" should reveal what you want, otherwise you could try "No autocompletion".
Anyway here's data for both the old location and new: [1]. There is some overlap because when you browse to a redirect, it will count as a pageview for the redirect itself, and not for the target article. If you want all pageviews, including redirects, consider using the "Redirect Views" tool: [2]
You probably are thinking to yourself that you always want redirects to be included, but doing so does not ensure completely accurate data. For instance, those redirects could have at some point in the past gone to a different article. So you'd then be getting (probably slightly) inaccurate numbers. There are talks of tackling this issue in the API itself, see [3]. In addition some pages have lots of redirects which could noticeably slow down the tool.
Another possible solution is to query the page move log. This is something I hope to actually do at some point, you can follow the ticket on Phabricator for updates [4]. This may be quite challenging, since there are some edge cases where the move log will steer you wrong and you end up querying the wrong pages.
Sorry for my excessive tech jargon, if it comes off that way :) Until we have a better way of handling this, first make sure you have the "Search method" set to "Autocomplete including redirects" or "No autocompletion", then simply type in the old page title.
Hope this helps,
~MA
[1] https://tools.wmflabs.org/pageviews/?project=fr. wikipedia.org&platform=all-access&agent=user&range=last- year&pages=Batman_v_Superman_:_L%27Aube_de_la_Justice| Batman_v_Superman_:_L%27Aube_de_la_justice https://tools.wmflabs.org/pageviews/?project=fr.wikipedia.org&platform=all-access&agent=user&range=last-year&pages=Batman_v_Superman_:_L%27Aube_de_la_Justice%7CBatman_v_Superman_:_L%27Aube_de_la_justice [2] https://tools.wmflabs.org/redirectviews/?project=fr. wikipedia.org&platform=all-access&agent=user&range=last- year&sort=views&direction=1&view=list&page=Batman_v_ Superman_:_L%27Aube_de_la_justice [3] https://phabricator.wikimedia.org/T121912 [4] https://phabricator.wikimedia.org/T141332
On Thu, Apr 13, 2017 at 11:10 AM, Joseph Allemandou < jallemandou@wikimedia.org> wrote:
Hi Gheorghe,
Thanks for your message and support :) The reason the page had no views before 4/29/2016is because its name has changed that day [1], and we collect pageviews by name.
You can find the raw data using direct API calls ([2] for new name, [3] for old name, only capital J differs ...), but I have not found a way to view it properly with the pageviews tool. Being able to request titles that are not current pages could be an interesting feature request -- Ping @MusikAnimal ;)
Cheers ! Joseph
[1] https://fr.wikipedia.org/w/index.php?title=Batman_v_Superman _%3A_L%27Aube_de_la_justice&type=revision&diff=125734639&oldid=125727682 [2] https://wikimedia.org/api/rest_v1/metrics/pageviews/per- article/fr.wikipedia/all-access/user/Batman_v_Superman_%3A_ L'Aube_de_la_justice/daily/2015010100/2016123100 [3] https://wikimedia.org/api/rest_v1/metrics/pageviews/per- article/fr.wikipedia/all-access/user/Batman_v_Superman_%3A_ L'Aube_de_la_Justice/daily/2015010100/2016123100
On Thu, Apr 13, 2017 at 4:10 PM, Gheorghe Postelnicu < gheorghe.postelnicu@gmail.com> wrote:
Hello,
First of all, thank you for providing such a wealth of information regarding Wikipedia usage. This is really interesting.
I was browsing the pageviews visualization provided by wmflabs and noticed a particular case which surprised. The FR page for *Batman v Superman *seems to have received no visit whasoever before 4/29/2016:
https://tools.wmflabs.org/pageviews/?project=fr.wikipedia.or g&platform=all-access&agent=user&range=last-year&pages=Batma n_v_Superman_:_L%27Aube_de_la_justice
However, the page existed as early as 2014:
https://fr.wikipedia.org/w/index.php?title=Batman_v_Superman _:_L%27Aube_de_la_justice&dir=prev&action=history
and similar pages in IT or ES show quite a lot of views:
https://tools.wmflabs.org/pageviews/?project=it.wikipedia.or g&platform=all-access&agent=user&range=last-year&pages=Batma n_v_Superman:_Dawn_of_Justice https://tools.wmflabs.org/pageviews/?project=es.wikipedia.or g&platform=all-access&agent=user&range=last-year&pages=Batma n_v_Superman:_Dawn_of_Justice
Any idea why the complete lack of pageviews for the FR page? Even if for some reason the page was not very popular in FR, shouldn't it have received at least some views?
Many thanks in advance for any ideas, Gheorghe
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics
-- *Joseph Allemandou* Data Engineer @ Wikimedia Foundation IRC: joal
Analytics mailing list Analytics@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/analytics