https://github.com/barzamin/fediverse-composition

https://github.com/barzamin/fediverse-composition

Last synced: about 13 hours ago
JSON representation

• Host: GitHub
• URL: https://github.com/barzamin/fediverse-composition
• Owner: barzamin
• Created: 2020-03-06T20:53:51.000Z (almost 5 years ago)
• Default Branch: master
• Last Pushed: 2020-03-06T21:03:40.000Z (almost 5 years ago)
• Last Synced: 2024-11-09T04:39:49.277Z (about 2 months ago)
• Language: Python
• Size: 1.95 KB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
something about [this analysis](https://rosenzweig.io/blog/the-federation-fallacy.html) strikes me as somewhat incorrect? and i think it's because distributions – if you average over all instances – will be skewed by two things:

1. instances have tons of dead unused accounts

2. being on an instance does not mean people listen to you

if you look at actual social relationships (or, rather, a shitty proxy for them in terms of followers/followees of a user), you see a much closer to pareto distribution of domains involved. this distribution is *different for each user*, based on their interests; this is why averaging across the whole fediverse is not a great way to figure out where people actually live.

social networks on the fediverse as anywhere else are relatively disjoint and as such form different constellations across servers. averaging across them obviously gives a skewed picture.

this isn't to stay fedi doesnt have a centralization problem, just that it's more complicated than looking at the distribution of user counts for all instances off of fediverse.network or something.