| debt | |
| Jun 30 2016, 8:34 PM |
| F4275324: Screen Shot 2016-07-14 at 3.38.24 PM.png | |
| Jul 14 2016, 10:39 PM |
| F4218045: Screen Shot 2016-06-30 at 10.28.08 PM.png | |
| Jun 30 2016, 8:34 PM |
| F4218044: Screen Shot 2016-06-30 at 10.28.41 PM.png | |
| Jun 30 2016, 8:34 PM |
Please investigate what might have happened on June 17th, 18th and 19th that caused a huge spike in the numbers reported for the 'no action' and 'search' results that are showing on the dashboard, as the numbers have not yet returned to 'normal' since.
| Subject | Repo | Branch | Lines +/- | |
|---|---|---|---|---|
| Revert commit from 20 June 2016 | wikimedia/discovery/golden | master | +7 -6 |
Started work -- acquired & refined data: https://github.com/wikimedia-research/Discovery-Research-Portal/commit/a5abfc59737067391e82796e86f84028e21dd9ab
Change 299085 had a related patch set uploaded (by Bearloga):
Revert commit from 20 June 2016
Done https://github.com/wikimedia-research/Discovery-Research-Portal/blob/master/Analyses/Drop%20in%20actions/report.pdf Sorry about the delay! Lots of technical problems + rendering took some time
One question on wording: "...snaply determined as..." in the first paragraph. Not sure that what means.
Also - are there any recommendations to how better get this data within our normal guidelines?
Thanks!
I wasn't sure how else to succinctly summarize "Well, I was looking at the code and that line stood out to me as unnecessary after all the discussions we've had about clickthrough rates and how they should be calculated, so I made a snap judgement to get rid of that line because it seemed counter to how we decided that metric should be calculated."
Also - are there any recommendations to how better get this data within our normal guidelines?
Could you please elaborate on what you mean?
LOL, fair enough. ;)
I was wondering which one of these options you wanted to try / go with, based on your analysis without having to do more logging of our visitor's actions/tracking:
"We conclude with a proposal of an alternative metric, either the “most commonly clicked section per visit” or “most common section clicked on across all visits per session”."
Also....the pageviews also has a spike during this same timeframe - is that ok or does it need to be adjusted?
It looks like the second one follows the original metric really closely, if not perfectly. The "most commonly clicked section per visit" metric is at least different enough. Hm.
Also....the pageviews also has a spike during this same timeframe - is that ok or does it need to be adjusted?
That one is completely independent and even excludes search-redirect.php! :D
