[guardian-dev] getting useful tracking data from our F-Droid repo without leaking privacy
tom at ritter.vg
Wed May 3 11:28:49 EDT 2017
FWIW, the Tor Project's log sanitization code is here:
https://gitweb.torproject.org/webstats.git/tree/src/sanitize.py and it
publishes the results publicly.
It rounds per-day.
On 3 May 2017 at 07:31, Hans-Christoph Steiner
<hans at guardianproject.info> wrote:
> Hey Nathan,
> I'd like to try to enable some privacy-preserving tracking on the
> Guardian Project F-Droid repo. Looks like logging was turned off August
> 19, 2014. What do you think about turning on the Apache logging again,
> but only keeping the logs for one day?
> For the F-Droid repo, there would be a cron'ed script that would take
> just what was downloaded and what time. I'm tempted to also convert the
> IP address to a country, and store that.
> Rounding off the time to the day seems like a nice balance of useful
> info without giving away too much. That eliminates the rich time-of-day
> metadata, e.g. night, morning, lunchtime, etc. But maybe rounding to
> the week would be better since that would also eliminate info about the
> weekly cycle (e.g. downloads on Friday evening are not likely to come
> from an orthodox Muslim, or Saturday for an orthodox Jew; then there are
> holidays, etc.).
> PGP fingerprint: EE66 20C7 136B 0D2C 456C 0A4D E9E2 8DEA 00AA 5556
> List info: https://lists.mayfirst.org/mailman/listinfo/guardian-dev
> To unsubscribe, email: guardian-dev-unsubscribe at lists.mayfirst.org
More information about the guardian-dev