The problem Base social graph solves

The Base ecosystem is rapidly growing, with numerous projects and influencers emerging. However, understanding the complex social dynamics and predicting trends within this community can be challenging using traditional metrics alone.

Solution

We've developed a tool that leverages network analysis to map and analyze the social fabric of the Base community on Farcaster. By treating the ecosystem as a dynamic social network, we can uncover deeper insights into community trends, project growth, and influencer impact.

Technical approach

Our data collection process focuses on Farcaster profiles associated with Base projects, influencers, and active community members. Now we mainly gather data on followers and followers.

Using this data, we construct a directional, dynamic graph representing the Base community on Farcaster. We then apply various network analysis algorithms, including Temporal PageRank, Community Detection, Influence Propagation Models, Network Growth Analysis.

Product

Our tool presents an interactive, visual representation of the Base Farcaster community. Users can explore:

• Project Growth Trajectories. Visualize how a project's network expands and evolves
• Influencer Maps. Identify key opinion leaders and track their impact
• Trend Prediction. Highlight emerging topics or projects based on changes in network structure and activity patterns
• Community Health Metrics. Assess the engagement and growth of different projects or sub-communities

Currently, users primarily interact with the dynamic graph to make their predictions and insights. In future versions, we plan to automate many of these analyses.

By analyzing these social dynamics, users can gain deeper insights into the Base ecosystem, potentially identifying promising projects early, understanding community sentiment, and making more informed decisions about where to focus their attention or resources within the Base community.

Challenges I ran into

• Accessing comprehensive Farcaster data through APIs proved challenging.

• Parsing Farcaster data for all Base projects and their networks takes forever. Had to implement multithreading to make it somewhat manageable, but it's still time-consuming.

• There's a ton of data to mine, clean, and analyze for each graph. The process is pretty chaotic and manual right now. Really need to automate this to scale the platform, but haven't had time to figure it out yet.