We used text to SQL to build a "Webflow for Web3 Dashboards". We discovered loads of problems in the entire blockchain data space outlined below.

Demo: Design your dashboard, type your query in English.

code: https://github.com/vatsalaggarwal/blockchain-text-to-sql prompt: https://llmhub.com/2/functions/23/share

We're open-sourcing everything.

There is A LOT of blockchain data (~1.5Tb on ETH+BTC). While this is "open data", it's basically unusable - there is too much for anyone to understand what's going on without tools that help humans make sense of it.

So, in October, we built out something hacky (the video is NOT cherry picked!!). Our aim was to make creating web3 dashboards like https://info.uniswap.org/#/ or https://dune.com/anngel/XDAO-Statistics as “no-code” as possible.

But, we found much deeper problems underneath. 3 types of problems: - liveness: time to access data that was last committed - quality: cleanliness/structure of tables - UI for business intelligence NOTHING WORKED.

We ignored liveness. On data quality, there were no public APIs we could use to get good tables! (we tried @flipsidecrypto but they took >1 day to return simple queries!). @DuneAnalytics had ridiculously good tables, but doesn't provide public access.

On UX/BI: most users either had to use SQL OR use prebuilt dashboards in Dune or Messari. Loads of nontechnical folks need this data, but can't use SQL and pre-made dashboards are insufficient!

In summary, the problems are: - Make good data available - Make it possible to run queries that return “current” results (i.e. not outdated) - Reduce the friction for non-technical users to convert their thoughts/questions into computer-executable queries.