Stop using Google Analytics

[choldgraf]

Context

We currently use Google Analytics to track launches on mybinder.org - however, recent rulings in Europe are concluding that Google Analytics violates GDPR (another ref from matomo. If this is true, then we should probably just stop using Google Analytics for now.

Alternatives?

I think it's important that we have some method of tracking which/how many repositories are launched, and ideally which countries those launches are coming from. This has been an important part of demonstrating Binder's impact across the world. There are some other alternatives out there:
,

• Matomo: We have discussed using Matomo for mybinder.org as well, and I seem to remember us setting something up. Is this working and is it a replacement of Google Analytics? I see mention of it in the docs but it's unclear where to access it (and the .yaml-based secrets sharing doesn't work for me right now, see Use sops instead of git-crypt + ssh-vault #473)
• Plausible: I've heard a lot of good things about Plausible, though it is a bit expensive. They have a self-hosted option but I'm not sure if we have the person bandwidth to maintain this.

Actions

• Agree whether we should turn off our Google Analytics tracking. I propose that we leave this open for a week to see if anybody objects or proposes an alternative approach. If nobody does, we should turn off Google Analytics.
• Turn off Google Analytics tracking via the web interface (I believe I have permissions to do this, happy to do so when I'm back from vacation)

[manics]

I agree we should disable Google Analytics.

[choldgraf]

As a side-note: I'd be happy to explore whether we can use grant funds to pay for Plausible, if that would help simplify our workflows. In my opinion, paying $99/mo for something is totally worth it if it means we don't have to spend ~any time thinking about maintaining it or providing access.

[minrk]

We're already using matomo on mybinder.org and have been for some time. I'm happy to turn off GA (I thought we already had! I've only looked at matomo for a while).

We anonymize data as much as GA allows, so I don't believe our use of GA does violate those issues (all mentions I see specifically reference tracking cookies and/or IPs-as-PII as relevant to the violation, but we anonymize IPs and disable GA cookies and respect DNT).

Our matomo server is operated in the US, which may be an issue? We also anonymize IPs htere, so there shouldn't be any PII governed by the data-transfer rules, but we do have tracking cookies enabled (by default) on matomo: jupyterhub/mybinder.org-deploy#2134

[choldgraf]

I believe that this is closed by jupyterhub/mybinder.org-deploy#2135 but somebody please correct me if that is wrong