I've been building data visualizations to understand editor behavior.
I'll be sharing a few of the interesting observations that have come out of them.
This is a work in progress :-)
Calculating retention is a bit tricky.
We are going to use a very simple definition.
All the active editors joining in a month will be considered the total in the group.
We will then look at what % of them are active in every subsequent month.
The groups are on the Y axis. Every row represents represents the activity of a group over its lifetime.
The X-axis shows the time in months
Filter by %
Filter by the % value of retention
Filtered at 5% - 100%
The activity of groups are stacked along the Y axis.
Time runs on the X axis
Filter by the age of a group in a month.
Now we have selected the contribution of only the new comers in a month.
Eg : The contribution of only the new comers in a month
Monthly activity graphs - de,fr,zh
Editor retention rates - en. We can pick declines & spikes in monthly activity by looking at the vertical streaks of green & red. The spike are caused when a lot of older cohorts suddenly get active & the declines are similarly when their activity falls.
The groups are on the Y-axis.
Article group retention - de
Time is on the X-axis
Article group retention for de filtered at 5%-100% levels. We see a fall for articles created after 2007.
Lets do more of data viz :-)
The graphs look very different when we define "edits >=100 as an active editor".