Real-time dashboard | Voters

Dune

Create

Home

Feature Requests

206

Real-time dashboard

lmatz

For important analytics, people may want a dashboard that is constantly updated in real-time.
Using the same SQL repeatedly to refresh the results may consume a lot of resources.
Streaming databases such as https://github.com/singularity-data/risingwave can incrementally compute the latest query results without rescanning the whole data again, and only do work that is porpotional to the size of data changes.
This may take real-time analytics on Dune to a whole new level.

August 19, 2022

Activity Feed

Sort by

Jon C

To add to this, although many of the popular dashboards on Dune offer a high-level view, I believe a streaming dashboard approach can offer high-granularity analytics with the ability to zoom in on more details.
This may be useful to trading, investment, regulatory/anti-crime, and MEV related outfits. For instance, the ability to zoom in on the top 1000 accounts on a group of DEXes/NFT market places and view an up-to-date, second-to-second aggregation of multi-timescale (e.g. 1m, 5m, 30m, 2h) and multi-level categories (DEX/NFT collection, chain, NFT artist, NFT platform) of their trading history, and to create alerts based on that. The ability to cluster related public keys and view their trading patterns across multiple chains in real time. The ability to create a "trending-on-twitter"-like functionality to provide a "live" experience in web3 products.
In addition, a streaming approach can provide the ability to "zoom out" to long-horizon historical views to do a historical analysis. Historical data can be compared against realtime data. This seems to be lacking in many of Dune's dashboards since the queries are always recomputed and the historical view is never stored anywhere. A streaming system that cheaply persists the historical windows can provide this richer experience at low cost.
Since the materialized views are always up to date, a platform like Dune can really benefit from reducing total cost of ownership due to shared tables. This is especially relevant to Dune which encourages users to build on each others' queries.