Week 6: Network Isolation
April 3, 2026
Hello everyone, and welcome back to my blog! This week, I analyzed network isolation across both Reddit and X communities. Network isolation measures how internally connected a community is compared to how much it connects with other communities. A higher isolation score means that most of the connections are within one community, which means there is a higher chance of the community being an echo chamber.
How it was calculated
To calculate network isolation, I used the graph structures I showed in my Week 3 blog to visualize the connections between communities. However, those graphs don’t show all the connections within the community, so I used a separate graph structure where each node was individual comments.
For each community, I calculated the number of edges that were inside one community versus the edges that connected to other communities. The equation I used to calculate the isolation score was:
Isolation = 1 – (external edges/total edges)
A score closer to 1 means that the community is more isolated, while a score closer to 0 means the community is quite connected to other communities.
Reddit Communities
Donald Trump Policies: 0.819634
Tax Policy: 0.780711
Economic Inequality: 0.748477
Healthcare Policy: 0.743123
Private Prisons: 0.733083
Abortion Access: 0.726770
Iran Sanctions: 0.718178
Religious Freedom: 0.685297
Middle East Policy: 0.660290
These results show that Reddit communities are well split into clusters. The highest isolation score is Donald Trump Policies with a score of 0.819634, meaning this community has the highest chance of being an echo chamber out of all the Reddit communities. Other communities like Tax Policy and Economic Inequality also had high isolation scores, meaning these communities also tend to act more like echo chambers. Even the lowest score from Middle East Policy at 0.660290, is still above 0.5, meaning that there are limited connections between that community and others.
X Communities
Political Campaigns and Elections: 0.911257
International Relations and Strategy: 0.910421
Media, Communication, and Public Discourse: 0.906196
National Security and Defense: 0.863192
Education and Teaching: 0.836330
Social Issues and Race Relations: 0.814220
Research and Scientific Development: 0.813017
Economic Policy and Finance: 0.782751
These scores show that communities on X are extremely isolated, with many of the scores being above 0.9. Political Campaigns and Elections and International Relations and Strategy both have scores above 0.91, meaning that many of the connections in these communities are internal. Media, Communication, and Public Discourse are closely behind, meaning that a lot of communication are very contained in specific groups. Even the lowest-scoring community, Economic Policy and Finance at 0.782751, has a high score. Overall, communities on X tend to be more isolated than those on Reddit.
Overall, communities on both Reddit and X are very isolated, but X showed higher isolation scores compared to Reddit, meaning echo chambers on that platform may be stronger.
In the following weeks, I will calculate the final results for the Echo Chamber Indices for all communities across both platforms, and I will be looking at what kind of metrics influence echo chambers on both platforms.
Thank you for reading, and I will see you all next week!
Harish
Reader Interactions
Comments
Leave a Reply
You must be logged in to post a comment.

This analysis looks good, just like last week’s. I was wondering, how far back does the dataset go? Would it be possible to do a pre vs post-Covid analysis, or a pre vs post-2016 analysis?
Great job this week! It’s pretty surprising (and a little concerning) to see how high the numbers are, especially on X, where so many communities seem to mostly interact within themselves. That really helps show how echo chambers can form without people even realizing it. I’m curious to see what factors you find contribute most to this in your next posts!
Hello Harish,
Great job! It’s interesting to see that X/Twitter is much higher. It also shows which social media sites (along with topic) tend to have higher inter-community discussions!
I’m curious about how you defined the “communities” on X compared to Reddit. On Reddit, communities are clearly partitioned into subreddits, but X is much more fluid. Did you use specific hashtags or keyword clusters to define those boundaries?