⚓ T119531 Analyse results of accept-language header test on or after 2015-12-17

Status	Assigned	Task
Resolved	mpopov	T119531 Analyse results of accept-language header test on or after 2015-12-17
Resolved	• Deskana	T119530 Verify data pipeline for accept-language test on or after 2015-12-10
Resolved	EBernhardson	T119529 Turn off A/B test for the accept-language test on or after 2015-12-17
Resolved	EBernhardson	T119528 Turn on A/B test for the accept-language test on 2015-12-10 after train deployment
Invalid	EBernhardson	T119527 Write A/B test for accept-language header, where the user is shown results from other wikis if they get nothing on their primary wiki
Resolved	EBernhardson	T120020 Write A/B test documentation for accept-language header test

• Deskana created this task.Nov 24 2015, 5:42 PM

• Deskana raised the priority of this task from to Needs Triage.

• Deskana updated the task description. (Show Details)

• Deskana added projects: Discovery-ARCHIVED, Discovery-Analysis (Current work).

Restricted Application added subscribers: StudiesWorld, Aklapper. · View Herald TranscriptNov 24 2015, 5:42 PM

• Deskana triaged this task as Medium priority.Nov 24 2015, 5:43 PM

• Deskana added a subtask: T119530: Verify data pipeline for accept-language test on or after 2015-12-10.

• Deskana closed subtask T119530: Verify data pipeline for accept-language test on or after 2015-12-10 as Invalid.Dec 1 2015, 5:39 PM

The test was actually turned on in November, so as of Dec 1 we have enough data to analyze. We would like to do at least an initial verification that the test data is valid as soon as possible.

• Deskana reopened subtask T119530: Verify data pipeline for accept-language test on or after 2015-12-10 as Open.Dec 1 2015, 6:02 PM

There was a mix-up; this test is not currently running like was thought when this tasks was closed. Altering dates back accordingly.

• Deskana renamed this task from Analyse results of accept-language header test on or after 2015-12-01 to Analyse results of accept-language header test on or after 2015-12-14.Dec 1 2015, 6:05 PM

• Deskana moved this task from Needs triage to On Sprint Board on the Discovery-ARCHIVED board.Dec 3 2015, 5:06 AM

mpopov mentioned this in T119530: Verify data pipeline for accept-language test on or after 2015-12-10.Dec 11 2015, 9:00 PM

mpopov added a subtask: T119529: Turn off A/B test for the accept-language test on or after 2015-12-17.Dec 14 2015, 5:54 PM

• Deskana renamed this task from Analyse results of accept-language header test on or after 2015-12-14 to Analyse results of accept-language header test on or after 2015-12-17.Dec 15 2015, 9:10 PM

Ironholds moved this task from Backlog to Stalled/Waiting on the Discovery-Analysis (Current work) board.Dec 16 2015, 9:09 PM

• Deskana moved this task from Stalled/Waiting to Backlog on the Discovery-Analysis (Current work) board.Dec 17 2015, 9:13 PM

• Deskana closed subtask T119530: Verify data pipeline for accept-language test on or after 2015-12-10 as Resolved.Dec 18 2015, 12:30 AM

mpopov moved this task from Backlog to In progress on the Discovery-Analysis (Current work) board.Dec 28 2015, 6:22 PM

mpopov claimed this task.Dec 28 2015, 10:12 PM

Draft in need of review before going on Commons: https://github.com/wikimedia-research/LangAcceptHeader/blob/master/report_results.pdf

mpopov moved this task from In progress to Needs review on the Discovery-Analysis (Current work) board.Jan 5 2016, 10:13 PM

"postulated" is unnecessarily academic.
The first sentence of the Introduction; will it, or is that our hypothesis? It shouldn't be made as a statement if it is.
Suggest linking to the last report inline, and in prose, rather than academic-style referencing.
What are status quo settings?
"pseudo-random yet deterministic" is too complicated. "The sampling process was random but on a per-user basis; if a user was selected, all of their queries were included"
IP/UA/XFF is not unique
Conclusion?