1500503 - Understand high devtools use in India and China

Status: RESOLVED FIXED

(Data Science :: Investigation, task)

Description

[Tim Smith (inactive) 👨‍🔬 [:tdsmith]]

We observed an excess number of Devtools panel opens in India and China vs the Firefox user base: https://sql.telemetry.mozilla.org/queries/52475#139970

Some previous examinations in the prerelease channel suggested that accidental use might be driving in India but not in China:

* https://metrics.mozilla.com/protected/tdsmith/20180508-devtools_by_country_draft.html

* https://metrics.mozilla.com/protected/tdsmith/20180626-onboarding_summary_reweighted.html

Are India and China really different? What do the release populations look like?

Comment 1

[Tim Smith (inactive) 👨‍🔬 [:tdsmith]]

In the release population, it seems plausible that both India and China just have a higher fraction of developers among people activating devtools.

* Users of devtools geolocated to China and India do seem to be more engaged than users from other locales, like Germany and the US.
* We should probably abandon the total-toolbox-open-count metric from the devtools dashboard; it’s easy to write a query for but not very illuminating.
* Users from India in the beta channel are not at all like users from India in the release channel. We should avoid experimentation in the beta channel.

https://metrics.mozilla.com/protected/tdsmith/2018-10-19-devtools_india_china.html

Comment 2

[Tim Smith (inactive) 👨‍🔬 [:tdsmith]]

Another look: the fraction of sessions where the toolbox active time is almost the entire session time is higher in China vs the US
https://dbc-caf9527b-e073.cloud.databricks.com/#notebook/41010/dashboard/41031