Research:First messages to new users

From Meta, a Wikimedia project coordination wiki
This page documents a completed research project.


Topic[edit]

Who is sending messages to new users, and what tools and scripts are they using? This sprint will break down these figures, particularly along the subjects of:

  • First messages sent to new users

Process[edit]

Using a live replicating copy of the English Wikipedia database, these types of first messages can be acquired through a few basic steps:

  1. Get every distinct user_name from enwiki.user
  2. Join enwiki.user.user_name with enwiki.page.page_title to get all user talk pages
  3. Join enwiki.revision.rev_page with enwiki.page.page_id to get all the first revisions to a user talk page
  4. Parse rev_comment to determine tool and message type (requires a list of all tools, their names, and the automated edit summaries they leave)
  5. Parse rev_user and user_id to determine metadata

Results and discussion[edit]

Proportional plot of first messages sent to users.

Raw plot of first messages sent to users.

Future work[edit]

Does the tool used and/or message type predict retention? Tricky, because tool users are selecting from different populations to begin with.