Steps to Reproduce: in the app, Spanish is describing cleopatra as trending. This is a default voice search test in spanish for hello google. It shouldnt be represented in the top articled for the week.
Actual Results:
Expected Results:
Description
Description
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Open | None | T280565 Improve pageview automated traffic detection heuristics | |||
| Open | BUG REPORT | None | T270784 App describing Spanish Wikipedia article "Cleopatra" as trending (due to default voice search test for Hello Google) |
Event Timeline
Comment Actions
Sounds like this is a test article whose readership is being artificially inflated by tests. (Presumably, those tests should not be included in reading stats.)
Not exactly sure which team this would lie with, so sending it upstream. Feel free to reassign further upstream if warranted.
Comment Actions
Does "upstream" mean "Wikifeeds"?
It's the same in the general pageview stats on https://pageviews.toolforge.org/?project=es.wikipedia.org&platform=all-access&agent=user&redirects=0&range=latest-20&pages=Cleopatra|Sorteo_Extraordinario_de_Navidad|Colegio_de_San_Ildefonso_(Madrid) so not sure there's anything to do in Wikifeeds itself. :)
Comment Actions
Upstream meaning "Product Infrastructure for now", but knowing that the source of this is probably upstream from them. (But we didn't know where it ultimately lies - hence encouraging them to send it further upstream as necessary.)
Comment Actions
Is there a way to figure out what are the articles used for testing? Maybe we can exclude them from wikifeeds. Other than that it doesn't look like an issue in the wikifeeds service level.
Comment Actions
Just bumping to say that users are still seeing Cleopatra in their top read.
https://es.wikipedia.org/api/rest_v1/feed/featured/2021/04/23
Comment Actions
Appears to be driving traffic on English now as well: https://pageviews.toolforge.org/?project=en.wikipedia.org&platform=all-access&agent=user&redirects=0&range=all-time&pages=Cleopatra
Comment Actions
This is a bit icky because the traffic is in fact genuine, as opposed to a disruptive bot or intentional manipulation of pageviews data. When we see spikes in traffic, we look for an explanation, and here we have one. Thus removing Pageviews-Anomaly. I agree we should still hide the article from the most-viewed section of the app and other places where we advertise trending topics, but removing it from the actual /top results in AQS is questionable. That data should reflect reality.