RE: What do Communities on Steemit Look Like? (Part 1)
Hi @eroche, great to see you contributing to Utopian as well! :)
Interesting approach to find communities based on voting activity! What were the criteria for the ML part to detect a community? Did it include only the number of votes, or also their value?
In the Utopian open source context it would have been great to have the results in a way that they can be reproduced. Could you give more details on the exact query, the machine learning part, and how the fancy graphs were created?
Unfortunately, only a few account names can be read in the graphs.
One name in the graph is chbartist, with frequent posts boosted into the hundreds of $ while having 5k own account SP. In this case it is not very surprising that the graph doesn't show many outgoing votes.
So I absolutely agree with your conclusion, that the voting activity as shown here is probably not the best measure to find communities. Looking forward to your next post!
Your contribution has been evaluated according to Utopian policies and guidelines, as well as a predefined set of questions pertaining to the category.
To view those questions and the relevant answers related to your post, click here.
Need help? Write a ticket on https://support.utopian.io/.
Chat with us on Discord.
[utopian-moderator]
Thanks @crokkon,
In this, my first attempt at this refining an algorithm to detect Steem communities based on network graphs, I have not looked at vote value directly but performed a clustering algorithm to detect modularity based on the density of connections between nodes. In this case based on Vote Count.
I was very interested to see if small groups were forming and if one or a small number of people was being rewarded for being part of community. That does not appear to be the case.
It would be interesting to extend this analysis to cluster based on a modularity calculated for Vote Amounts, do big accounts form communities that vote for each other?
In my next contribution where I will focus on comments, I will go into more detail about the tools and methodology used for this analysis.
Thank you for your review, @crokkon!
So far this week you've reviewed 1 contributions. Keep up the good work!