Page MenuHomePhabricator

Provide metrics for WMF quarterly report about April-June 2015
Closed, ResolvedPublic

Description

Provide the below numbers for the quarterly report (publication in final form due July 30, see T106501), in the same format as for last quarter's scorecard). Compare the analogous ticket from last quarter: T97344

I am creating separate blocking tasks for some of these metrics as I go along, and will link them in the below list.

Participation

Site reliability

  • Uptime (will probably use the new Catchpoint metrics for readers and contributors given here, see also here)
  • Read latency: T106507
  • Write latency: T106508

Readership

  • Page Views (currently blocked on T105107: Restart Pentaho)
  • Visitors (will need to include the legacy comScore number again)

Content

  • New articles during April-June 2015 (i.e. article increase from March 31 to June 30) - will probably use Emausbot's record on Meta again
  • Edits (needs update of https://reportcard.wmflabs.org/graphs/edits with June data)

Fundraising metrics (from FR team):

  • Amount raised
  • % raised on mobile

The scorecard will still be marked as "beta" in the quarterly report, since this selection of metrics may be revisited in coming quarters.

Event Timeline

Tbayer raised the priority of this task from to Needs Triage.
Tbayer updated the task description. (Show Details)
Tbayer subscribed.
Krenair subscribed.

It looks like you know what you're doing here, so I'm just going to assign it to you and ignore the lack of projects listed

It looks like you know what you're doing here, so I'm just going to assign it to you and ignore the lack of projects listed

Thanks @Krenair! Help with sorting out the Phabricator bureaucracy is welcome, but yes, the lack of project did not significantly impede the completion of this task. Last quarter I put it into the Analytics and WMF-Product-Strategy projects, but the latter is no longer active (and we haven't set up one for COO related areas). Feel free to add this one to Analytics retroactively (I tried to separate out all tasks that needed support from the Analytics team, but I guess the Analytics workboard's scope is by topic rather than by team).

All the data was obtained in time and published on July 30 as part of the Q4 report (slide 3). Thanks to everyone involved!

To record some notes, in particular regarding calculation of the quarterly averages and trends:

Articles:
I continued to use EmausBot's records (see also T97476), assuming as last quarter, per discussions on Meta, that the recent article count corrections did not have a large effect on the global numbers.

35 446 007 at 00:00, 1 July 2015 (UTC)
34 843 225 at 00:00, 1 April 2015 (UTC)
34 127 177 on 00:00, 1 January 2015 (UTC)
32 202 860 on 00:00, 1 July 2014 (UTC)
31 176 373 on 00:00, 1 April 2014 (UTC)

That's (35 446 007 - 34 843 225 ) / 91 ~ 6624 articles per day in Q4,
and about 14.9% less than in Q3:
(((35 446 007 - 34 843 225) / 91) / ((34 843 225 - 34 127 177) / 92)) - 1 = -0.148

and a 41.3% year-over-year decline:
(((35 446 007 - 34 843 225) / 91) / ((32 202 860 - 31 176 373) / 91)) - 1 = -0.4127

Signups
See here again for the methodology and the complete historical numbers. There were some small correction for past months because a few recently created projects had been accidentally omitted earlier.

Total number for Q4: 713 884 + 195 341 + 419 895 = 1329120

compare to Q3:
((713 884 + 195 341 + 419 895) / (514 498 + 477 897 + 514 508)) - 1 = -0.1179...

year-over-year:
((713 884 + 195 341 + 419 895) / (320 647 + 341 800 + 310 818)) - 1 = 0.3656...

(BTW, obviously it's possible that this show some SUL finalization and mobile signup effects.)

Pageviews
Per discussion at T105107: Restart Pentaho we decided to get these directly from Hive instead of having to restart Pentaho. See also https://etherpad.wikimedia.org/p/analytics-notes . Unfortunately it turned out there appear to be some small but non-neglible differences between the new unsampled data on Hive, and the sampled historical data used for the previous quarterly report. I have followed up with Kevin and Dan to see if we can resolve or explain these, but had to leave out the trend data for this one at publication time.