Maniphest T196785

Confusing results in Turnilo
Closed, ResolvedPublic
Actions

Assigned To

Authored By

	MMiller_WMF
	Jun 8 2018, 10:51 PM

Description

I think the best person to tag here is @Ottomata.

I'm looking at some data in Turnilo, and in particular looking into a spike in pageviews in France on May 20, in which pageviews doubled.

When I split this out by Agent Type, I can see that it looks like the spike in traffic is from spiders. Makes sense.

But then when I split it out by Project, the spike is totally gone. Though Turnilo only lets me display ten lines on the graph, I've selected the top 10 most popular wikis in France, and still seeing no spike.

Do you know where that spike is hiding when I split by project?

Event Timeline

MMiller_WMF created this task.Jun 8 2018, 10:51 PM

Restricted Application added a subscriber: Aklapper. · View Herald TranscriptJun 8 2018, 10:51 PM

See: https://bit.ly/2sJsyD7

Spike is due to a bot (python-urllib), best visible if you split by UA: https://bit.ly/2HB6LlA and filtering by that bot you can see the projects selected (wikisource and incubator) are the ones most affected, those are small and not displaying on your top list.

• fdans triaged this task as High priority.Jun 11 2018, 4:04 PM

• fdans added a project: Analytics-Kanban.

• fdans moved this task from Incoming to Data Quality on the Analytics board.

Nuria claimed this task.Jun 12 2018, 6:15 PM

Nuria moved this task from Next Up to Paused on the Analytics-Kanban board.Jun 18 2018, 3:05 PM

• Vvjjkkii renamed this task from Confusing results in Turnilo to 2cbaaaaaaa.Jul 1 2018, 1:05 AM

• Vvjjkkii removed Nuria as the assignee of this task.