Bringing data science into blockchain
summary
Concrete metrics and pipelines get recommended: compute total transfer volume (native token and USD), transaction count, active/new address growth, and transaction latency (XRP ~3s, Ethereum 15s–1 day). Build ETLs to parse raw ledgers, store fact/dimension tables for analytics and a graph database or directed graph to detect churn and cycles. Classify addresses (exchanges/payment providers ~300–400 self‑reported) via features and semi‑supervised models; use The Graph for indexing when available.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 1:02Blockchain foundations
- 3:24The problem of trust
- 6:39Blockchain mechanics
- 8:44Data science metrics
- 16:12Peer-to-peer utility
- 23:24Data architecture
- 27:04Future opportunities
- 29:12Q&A session
related talks
26mEthereum Blockchain Analytics
31mBlockchain + Tracking = Awesome Analytics
36mDecrypting Tech Hype for the Busy Coder
44mSolving real-world problems with distributed ledger technology
24mState of Blockchain
19mScaling Blockchain Games: The Fundamentals
13mThe End of Old Money
18mEnergy Traceability via Open Source Blockchain: A Comprehensive Solution for 24/7 Matching...