Post by Onah Innocent
Blockchain Data Engineer | SQL, Python
TradesLens just shipped a live dashboard and the story behind it is the pipeline itself. From raw on-chain events to a fully interactive DEX terminal, every number visible on screen flows through a data pipeline I designed and built from scratch. No third-party data. No external APIs. Just clean, end-to-end data engineering. 🔗 https://lnkd.in/d3QNzV-H Here's how the stack works: Ingestion : Envio HyperIndex listens to swap events across Uniswap V3, SushiSwap V3, and Solidly V3 on Ethereum, Arbitrum, and Optimism. Raw on-chain data lands directly into TimescaleDB. Transformation: dbt takes over from there. Staging models clean and normalize the raw data. Intermediate models apply protocol-specific logic per DEX. A single fact table unifies everything across protocols. Incremental builds only, triggered automatically via GitHub Actions every 2 hours. API Layer : The Gold layer feeds a FastAPI backend built with Python and SQLAlchemy, running time-series optimized queries on Postgres + TimescaleDB. Clean API contract between the data and the frontend , no raw DB queries on the client side. Dashboard: React 19 + Vite. A DEX terminal UI with real-time charts that speaks only to the API. What you see on screen is a direct reflection of the data engineering process underneath it. What makes this interesting isn't just the dashboard. It's that every data point on screen traces back to a single pipeline ingestion, transformation, API, display. End-to-end. no external dependencies. The full stack is open source would love feedback and contributions from anyone building in DeFi data, on-chain analytics, or data pipelines in general. Stack: Envio · TimescaleDB · dbt · GitHub Actions · FastAPI · React 19 🔗 https://lnkd.in/dengCQcy