Research talk:Editor re-activation

From Meta, a Wikimedia project coordination wiki

Editor death and reactivation period[edit]

The frequency of time between edits with EM model fits of log-normal distributions.

So, I did some research with edit sessions. In order to draw a temporal cutoff that would determine when a session was over, I built a model to explain the time between edits. It turns out that three clusters explained the distribution the best. Here's how I describe them:

  1. When edits are saved within minutes to an hour of each other, an editor is "within session"
  2. When there's a gap between edits of an hour to a month, an editor is "between sessions"
  3. When there's a gap between edits of a month to a year, an editor is "on a break"

I suspect that the third cluster represents editor reactivation and that this empirical investigation provides a good way to identify appropriate cutoffs for measuring reactivation rates. Thoughts? --Halfak (WMF) (talk) 16:37, 3 February 2014 (UTC)[reply]