Applying anonymized usage data to help the Singer and Meltano communities

[aaronsteers]

Opening this issue to discuss use cases where data from anonymized usage stats might help strengthen the community and provide valuable insights to community members and new users of Meltano and/or Singer.

RE: data privacy

Note: All of the data we collect is carefully anonymized. We go out of our way to not collect any user-entered data in clear-text. Even pip_url is hashed using a one-way hash, so we can only interpret the value if we already have the value as a known entity in our database.

[aaronsteers]

Just to kick off this discussion:

As of today, we use anonymized usage stats for the following:

• We surface to users how various forks compare against each other in terms of overall usage, number of projects installed, etc.
• We compare "project installed" vs "projects run successfully" per plugin, in order to determine where problem areas exist.
• When problem areas are found - for instance, 200 projects installed vs 2 project run successfully, we will (soon) update the maintenance status and/or present other warning messages to prospective users. We may also reach out to maintainers to see if we can assist; sometimes these are driven by documentation issues or other onboarding challenges.

[visch]

(#818 (reply in thread))