after my @JupyterCon keynote yesterday, @jason_grout brought up the activity plateau in this chart, where total activity across Jupyter projects stopped growing in 2016 after exponential growth up to that point

Counting *unique* contributors, numbers are still growing (2020 is a rough year, plus it's not over yet), so what gives?

(Huge caveat: to first order, these totals are mostly counting GitHub comments, a _very_ rough proxy for contribution activity, and loads of activities are not represented here)

The vast majority of lifetime contributors across Jupyter projects have only contributed a handful of times. ~90% of folks less than 10 times.

If we bin contributions into the most active 5, 10, etc. people any given year, we can see that contributions from outside the top 100 are still growing steadily. So where does the plateau come from?

Reversing the order, we can see that the 5 most active folks (full-time contributors) have dropped considerably, contributing about half in 2019 what they did in 2016. The top 10 in 2019 still don't add up to what 5 folks contributed in 2016.

(speaking as one of the five in 2016, that was not a sustainable level of activity, so I’m glad we aren’t doing that anymore!)

I think we can conclude that the @ProjectJupyer community is still growing, but we are struggling to spread out our maintenance responsibilities to keep up, and as an organization we are still vulnerable to the burnout / changing careers of a pretty small number of people.

Notebook with data and analysis here: https://nbviewer.jupyter.org/github/minrk/jupyter-activity-data/blob/main/plateau.ipynb

and try it yourself on @mybinderteam of course: https://mybinder.org/v2/gh/minrk/jupyter-activity-data/main