@Kgraessle @DMburugu - Have the events needed for this dashboard been instrumented? Or is that still in progress?
Once the instrumentation is complete, I can start to set up the dashboard with the metrics identified in the task description.

Hi @MNeisler

Nope, they are still in progress/code review.
Though I realize this ticket is missing two parent tasks so I went ahead and updated it (hopefully that makes it a little clearer where we're at with instrumenting the events).

Once the instrumentation is complete, I can start to set up the dashboard with the metrics identified in the task description.

Sounds good, thanks for the update.

@MNeisler let us know if you need specific help, just watching this on the radar for now.

In T412137#11550018, @MNeisler wrote:

@Kgraessle @DMburugu - Have the events needed for this dashboard been instrumented? Or is that still in progress?
Once the instrumentation is complete, I can start to set up the dashboard with the metrics identified in the task description.

We have the instrumentation up and running on test wiki; this should be ready for you @MNeisler

An initial draft of the superset dashboard is complete and ready for review. Please let me know if you have any questions or suggestions.

This dashboard includes aggregated data for the following metrics:

• Unique dashboard visitors. Total and daily rates.
• Second Week Retention Rate.
• Clickthrough Rates (CTR) for each link available in the personal dashboard. I included the following two types of CTR:
  • Proportion of users that clicked link at least once (CTR Per User):
  • Proportion of page visits that resulted in a click (CTR Per Page Visit):

Note: Data currently reflects all events logged to date including events logged on testwiki; however, I'll plan to filter out testwiki so we can limit the Dashboard to reflect pilot wiki data. This can also currently be done by selecting only the pilot wikis using the wiki filter.

A Superset dashboard which shows the rate in which editors who visit the Dashboard make edits, undo actions, send thanks, review edits, and patrol edits

If ok, I recommend this be done as a standalone pre/post deployment analysis rather than a dashboard metric. Including this on the dashboard would likely result in timeout errors due to joins across large datasets.

Instead, I can complete a separate analysis in a notebook to compare the rates at which users performed each of these actions before and after visiting the dashboard. A standalone analysis is better suited for this type of metric and would allow for more granular exploration, such as segmenting by user edit counts.

In T412137#11690460, @MNeisler wrote:

An initial draft of the superset dashboard is complete and ready for review. Please let me know if you have any questions or suggestions.

This dashboard includes aggregated data for the following metrics:

• Unique dashboard visitors. Total and daily rates.
• Second Week Retention Rate.
• Clickthrough Rates (CTR) for each link available in the personal dashboard. I included the following two types of CTR:
  • Proportion of users that clicked link at least once (CTR Per User):
  • Proportion of page visits that resulted in a click (CTR Per Page Visit):

Note: Data currently reflects all events logged to date including events logged on testwiki; however, I'll plan to filter out testwiki so we can limit the Dashboard to reflect pilot wiki data. This can also currently be done by selecting only the pilot wikis using the wiki filter.

Thanks, this looks great!

A Superset dashboard which shows the rate in which editors who visit the Dashboard make edits, undo actions, send thanks, review edits, and patrol edits

If ok, I recommend this be done as a standalone pre/post deployment analysis rather than a dashboard metric. Including this on the dashboard would likely result in timeout errors due to joins across large datasets.

Instead, I can complete a separate analysis in a notebook to compare the rates at which users performed each of these actions before and after visiting the dashboard. A standalone analysis is better suited for this type of metric and would allow for more granular exploration, such as segmenting by user edit counts.

NOTE: To use this dashboard you need to be a member of the analytics-privatedata-users group. If you are receiving errors about permissions, you will need to request SSH-less membership to this group. See wikitech:Data Platform/Data access for instructions on requesting

That makes sense to me, especially if it could potentially cause timeouts.

@Samwalton9-WMF What do you think, is that ok to do the last portion A Superset dashboard which shows the rate in which editors who visit the Dashboard make edits, undo actions, send thanks, review edits, and patrol edits as a pre/post deployment analysis?

Yeah that totally makes sense to me.

I've made the following updates to the superset dashboard:

• Updated the second week retention rate table to show per wiki breakdowns
• Added a time filter to view metrics over a selected time range.

Let me know if you have any questions or issues loading/filtering the tables.

cc @Samwalton9-WMF @Kgraessle