Livepeer's network data is fragmented across isolated systems, making it impossible to answer basic questions about network performance, AI model adoption, or orchestrator health. This hackathon project takes the first step toward democratizing network analytics by building a unified data pipeline that anyone can query.
<aside> 🎯
What We're Building: A proof-of-concept pipeline using StreamR → ClickHouse → Metabase that demonstrates how network-wide data can be collected, stored, and made accessible to the entire Livepeer community.
</aside>
This isn't just about better dashboards—it's the foundation for the Livepeer Data Lake, where orchestrators, gateways, developers, and researchers can all access the insights they need to optimize the network.
<aside> ⚠️
Our core problem is that answering simple questions like: "How many AI jobs have been run in the last 24 hours for each model?" is impossible at a network-level.
</aside>
Our network data is scattered across isolated islands:
We're flying blind on fundamental network questions: