What should a live economic signal feel like?
A live dataset can still feel dead if the interface treats it like a warehouse.
There is a strange product problem in data: the information can be live while the interface feels archival.
Rows, filters and dashboards are excellent once you know what you are looking for. They are less good at communicating the first-order fact that something in the world is changing now.
We have been thinking about Evericade less as a dashboard and more as an instrument. An instrument should show movement, baseline and change before it asks you to configure a report.
That does not mean animation for its own sake. A blinking screen is not more real-time than a static one. The interface should move only when movement carries meaning: a relationship accelerates, a new cluster forms, an observation arrives, a baseline breaks.
This matters because the first experience of a new data category is partly conceptual. Before somebody asks for an export, they need to understand what the thing is.
The best version of the product may therefore have two personalities: an instrument for seeing what is moving, and infrastructure for interrogating the underlying graph.
If those two layers work, the interface should make the dataset feel less like a report about companies and more like a window onto work happening between them.