Overview

The project focused on understanding how communication organizes itself within an informal online community. Using Discord interaction data, I explored who the most active users were, how users clustered around shared interests, and whether the network showed evidence of a core-periphery structure.

 

Problem

Online communities generate large amounts of interaction data, but communication patterns are often difficult to interpret without structural analysis. This project examined:

​

• Which users were most central to communication
• How user interests and demographics related to interaction patterns
• Whether communication was evenly distributed or concentrated among a smaller core group
• How tenure and activity influenced network position

 

My Role

I independently conducted all stages of the project, including:

​

• Discord data collection and processing
• Survey design and node attribute collection
• Full message history extraction
• AI-assisted and manual edge list construction
• Network visualization and statistical analysis in R using igraph, tidygraph, and ggraph
• Interpretation of network theory concepts including degree centrality, homophily, and core-periphery structure

 

Key Insights

• Communication was concentrated among a smaller set of highly connected users
• The server displayed a visible core-periphery structure rather than evenly distributed interaction
• Longer-tenured users generally appeared more central within the network
• Most users fell within a narrow college-age demographic range
• Some evidence of homophily appeared around shared interests such as gaming and social interaction

 

Solution

I created a weighted communication network by combining:

​

• Survey-reported communication patterns
• AI-detected direct reply interactions
• Manual review of contextual conversations and mentions
• Node attributes including age, tenure, activity level, and interests

​

The resulting visualizations highlighted user centrality, tie strength, demographic trends, and network structure through multiple graph and distribution analyses.

 

Impact

The project demonstrated how social network analysis can reveal meaningful communication structures even within small informal online communities. The analysis showed that interaction patterns were highly structured rather than random and that a small number of users drove much of the server’s communication activity.

​

Documentation

• Full Research Report
• Network Visualizations
• Edge & Node Attribute Datasets
• R code to generate visualizations
• Survey Methodology & Data Processing Notes