81 releases (12 breaking)

new 0.14.0-alpha.2 Feb 19, 2024
0.13.0-alpha.3 Jan 29, 2024
0.12.0-alpha.2 Dec 26, 2023
0.11.0 Nov 28, 2023
0.4.0 Mar 28, 2023

#73 in Machine learning

Download history 460/week @ 2023-11-02 668/week @ 2023-11-09 348/week @ 2023-11-16 666/week @ 2023-11-23 400/week @ 2023-11-30 755/week @ 2023-12-07 864/week @ 2023-12-14 603/week @ 2023-12-21 210/week @ 2023-12-28 1133/week @ 2024-01-04 641/week @ 2024-01-11 520/week @ 2024-01-18 1176/week @ 2024-01-25 1066/week @ 2024-02-01 987/week @ 2024-02-08 1543/week @ 2024-02-15

4,881 downloads per month
Used in 5 crates (4 directly)

MIT/Apache

100KB
2K SLoC

Rerun's analytics SDK

Part of the rerun family of crates.

Latest version Documentation MIT Apache

We use PostHog to collect anonymous usage statistics.

Usage data we do collect will be sent to and stored in servers within the EU.

You can audit the actual data being sent out by inspecting the Rerun data directory directly. Find out its location by running rerun analytics config.

Opting out

Run rerun analytics disable to opt out of all usage data collection.

What data is collected?

The exact set of analytics events and parameters can be found here: https://github.com/rerun-io/rerun/blob/main/crates/re_viewer/src/viewer_analytics.rs

  • We collect high level events about the usage of the Rerun Viewer. For example:
    • The event 'Viewer Opened' helps us estimate how often Rerun is used.
    • The event 'Data Source Connected' helps us understand if users tend to use live data sources or recordings most, which helps us prioritize features.
  • We associate events with:
    • Metadata about the Rerun build (version, target platform, etc).
    • A persistent random id that is used to associate events from multiple sessions together. To regenerate it run rerun analytics clear.
  • We may associate these events with a hashed application_id and recording_id, so that we can understand if users are more likely to look at few applications often, or tend to use Rerun for many temporary scripts. Again, this helps us prioritize.
  • We may for instance add events that help us understand how well the auto-layout works.

What data is NOT collected?

  • No Personally Identifiable Information, such as user name or IP address, is collected.
    • This assumes you don't manually and explicitly associate your email with the analytics events using the analytics helper cli. (Don't do this, it's just meant for internal use for the Rerun team.)
  • No user data logged to Rerun is collected.
    • In some cases we collect secure hashes of user provided names (e.g. application_id), but take great care do this only when we have a clear understanding of why it's needed and it won't risk leaking anything potentially proprietary.

Why do we collect data?

  • To improve the Rerun open source library.

Dependencies

~9–19MB
~250K SLoC