Page MenuHomePhabricator

Track pageview stats for outreach.wikimedia.org
Closed, DuplicatePublic

Description

Samir would like to use the pageview API to get stats for Outreach wiki, but currently those stats don't make it from wmf.webrequest to wmf.pageview_hourly. I remember some wikis were purposefully excluded but I don't see that rationale applying to Outreach. Let's triage at the next meeting.

Event Timeline

Milimetric raised the priority of this task from to Needs Triage.
Milimetric updated the task description. (Show Details)
Milimetric added a project: Analytics-Backlog.

In the pageview definition this domain is excluded on purpose at research's team request.
Since pageview_hourly only stores what we consider pageviews as per pageview definition: https://meta.wikimedia.org/wiki/Research:Page_view

we might need to add requests from this domain (and others) to a different table that is as easy to query as pageview_hourly is.

cc: @Ironholds

Milimetric triaged this task as Medium priority.Nov 19 2015, 6:16 PM
Milimetric moved this task from Incoming to Backlog on the Analytics-Backlog board.
Milimetric set Security to None.

+1000 please make this happen!

@Nuria: We'd really like to see this prioritized.

Pageviews are one of the metrics which many WMF teams use as markers of impact and measures of success. [1] Yet our team cannot because our home wiki is not considered by some to be an important enough part of the ecosystem.

Nevermind that data from webstatscollector (apparently no longer available at wmflabs) showed that Outreach wiki has a much more active community than, for example, Foundation wiki does. [2]

In fact... Outreach wiki has on average:

  • 3 times the number of namespace edits,
  • 4 times the number of active editors,
  • 5 times the number of accounts created, and
  • +1000% more pages created.

Furthermore:
--For the past 1 year, Outreach wiki has averaged around 8,000 daily pageviews. (per webstatscollector)
--For the past 6 months, Outreach wiki has averaged around 10,000 daily pageviews. (per webstatscollector)

Outreach wiki supports the Movement's 3 core programs (GLAM, Libraries, and Education) -- programs which consistently demonstrate impact on and add value to the Wikimedia projects. And it feels really unfair that the Outreach community has to beg for stats which all other wikis take for granted.

@Milimetric told me by email that "outreach wiki is not considered "content" by the research team and so the pageview definition they wrote specifically excludes outreach wiki data from all pageview data and APIs."

To that, I ask: Why, then, isn't Meta wiki also excluded on the same grounds? After all, that content is similarly not 'content' either. The type of content created by contributors on Outreach wiki mainly is pretty much a paraphrase of the first sentence of our mission statement. [3] Outreach wiki is the 'how' helps the 'what'! :)

Thank you for considering this request. And please see these related Village Pump posts. [4] [5]

[1] https://meta.wikimedia.org/wiki/File:Wikimedia_Foundation_Quarterly_Report,_FY_2015-16_Q1_(July-September).pdf
[2] https://outreach.wikimedia.org/wiki/Wikimedia:Village_pump/Archive_3#Proposal:_merge_the_Outreach_wiki_into_the_Meta_wiki
[3] https://wikimediafoundation.org/wiki/Mission_statement
[4] https://outreach.wikimedia.org/wiki/Wikimedia:Village_pump#More_page_stats_needed
[5] https://outreach.wikimedia.org/wiki/Wikimedia:Village_pump#Another_stats_question:_total_pageviews

Nuria renamed this task from Track stats for outreach.wikimedia.org in pageview_hourly to Track pageview stats for outreach.wikimedia.org .Dec 22 2015, 9:35 PM

@AKoval_WMF:
There are many, many systems we have whose requests do not make it into pageview_hourly but in this case that doesn't seem to be the core issue. If I understand correctly you want to have an easy way to count pageviews to outreach wiki, as easy as a select on pageview_hourly would be. Correct?

@Nuria Thanks for hearing me out. I do not know how easy it is to select on pageview_hourly, so I cannot answer that question. But, yes, the Outreach community would appreciate an easier way to count pageviews stats, please.

But it's not only pageviews stats we need, it's just more stats in general. [0]

[0] https://outreach.wikimedia.org/wiki/Wikimedia:Village_pump#More_page_stats_needed

Thanks for the supporting comment @AKoval_WMF and for your reply @Nuria!

Yes, Nuria, that's right. Any tool that we can use to easily count pageviews would do the job. I used to spend hours to collect the pageviews for some pages in a specific period of time. I was surprised by the API but unfortunately it isn't deployed on Meta!

Adding my +1 for pageviews on outreach to be reconsidered. @AKoval_WMF puts it better than I can, but it would be a huge help for our team (since it's our home wiki) as well as our colleagues in GLAM and Libraries, I'm sure.

@AKoval_WMF, @TFlanagan-WMF: are you aware you can get pageviews for outreach for the last two months right now? there is no additional work needed. You just need to have a developer /analyst with access to the cluster execute a simple select to get those numbers.

Thanks, @Nuria. Is the process you mention quick and easy? I'm just thinking ahead if we need to report some pageview numbers for internal or external reporting. I think we'd like an option to sift through the numbers ourselves, since we don't have a developer or analyst on our team.

Copying/mentioning @Selsharbaty-WMF since he has collected these numbers for our team in the past, and started this conversation about Outreach.

Thanks, @Nuria. Is the process you mention quick and easy? I'm just thinking ahead if we need to report some pageview numbers for internal or external reporting. I think we'd like an option to sift through the numbers ourselves, since we don't have a developer or analyst on our team.

Copying/mentioning @Selsharbaty-WMF since he has collected these numbers for our team in the past, and started this conversation about Outreach.

The process is not super duper completely easy, but it's definitely not hard. I'd be happy to sit down with someone and show them around the data.