Wikimedia Discovery/Meetings/Analysis retrospective 2017-03-03



Note that "mad" and "sad" don't have to mean literally angry or saddened. They can be used in a playful way as well.

Previous action itemsEdit

  • Dan: Chelsy/Mikhail should be optional at search sprint planning meeting    Done

What has happened since the last retro (2016-12-16)Edit

(Initial list taken from Discovery/Status updates)

  • Added cool new feature to the Metrics dashboard that highlight the selected date range in the sparklines (Go Chelsy! :D)Thanks!
  • Added a full geographic breakdown to the Portal dashboard to see traffic and clickthrough rates on portal across all countries
  • Added final results of the successful long term testing of the addition of app and app page links to portal page
  • Fixed a data bug where there was a spike in user engagement
  • Finalized the second BM25 testing analysis
  • "Hiring a data scientist", a post on Wikimedia Blog that provides an in-depth look into Discovery's hiring process for the Analysis team.
  • Wrapping up migrating a significant amount of data using the ReportUpdater infrastructure - almost done! Updating dashboards now to use the new datasets, including NEW datasets (like LDF endpoint usage for WDQS)
  • Inter-wiki search is here! An A/B test of a new search results page which includes inter-wiki search results was deployed to Persian, Italian, Catalan and Polish Wikipedias. Take a look and give us feedback! (T149806)

What has made you mad?Edit

  • Something going on with full-text search API counting in the new, Reportupdater-based data collection codebase (New, Chelsy is investigating, no ticket needed because it's still part of the migration/patch CR)
  • Reportupdater migration has taken up so much time from other, more fun projects

What has made you sad?Edit

  • Updating to reportupdater seemed harder and took longer than anticipated because of bugs, etc - no blame here - just noting that it took longer than we had hoped
    • We probably could have finish it faster, but both Mikhail and I are really tired of fixing those non-stopping bugs... Not a very fun project... :(
    • But needed to be done :)
  • Maps tile usage in Pivot - because of *things* has now been removed, maybe it'll be picked up again later on
  • Dan's recent absences +1+1+1
  • Can sometimes feel isolated working in a small team, with limited number of people just within Discovery

What has made you glad?Edit

  • Yay for blog posts!
    • And positive feedback from folx in the org + datasci community+1+1
    • Working with Melody K. was great -- she's a terrific resource
  • Chelsy does great work as co-analyst, very glad to be working with her -- MP Yay! You too!+1
  • Deb does excellent work as Product Manager for the team Yay! +1 yay - thanks!!
  • Things seem to be ticking along just fine with Dan's absence, which is definitely good that I'm not a single point of failure
  • - this is a perfect example of how things should work (on both ends), closing the task generates discussion, priorities are reassessed, task is reopened. "Easy come, easy go."+1
  • Highlighting the work that the Analysis team has done was noted and praised in the Quarterly Review meeting (Jan 2017)

What has made you plaid? (aka whatever didn't fit in one of the other categories)Edit

  • Deb was sleepless in Taiwan ;)
    • Is there a Tom Hanks and Meg Ryan movie coming out about that?
  • Dan was jet lagged, is jet lagged, and will be jet lagged
  • Do other teams in Discovery feel like we are supporting them enough?


Do other teams in Discovery feel like we are supporting them enough?Edit

  • Search team seems to work well. Good communication. 
  • Seems less true with maps, but some of that is historical; right now there isn't much analysis to do
  • In October, did a WDQS report, and there is a dashboard, so hopefully Stas feels supported
    • Yes, he has spoken positively. He would speak up if he's missing something, and hasn't
  • There was a question about reporting on graphs usage / long term maintenance (who's using them)
    • Maybe not an analysis thing--probably just some queries

Can sometimes feel isolatedEdit

  • Consider reaching out to other teams to see if they could use some help. 10% time is available, if nothing else
  • Flexibility to go above 10% depending on current Discovery workload

Questions about this retrospectiveEdit

  • What is the best frequency?
    • Still monthly
  • Ready to try a different format?
    • Some support for keeping this one

Action itemsEdit

  • (none)