Registry

Two registry tables describe who runs what. Both come from the registry we author by hand, so a row exists because somebody wrote it down, not because we detected activity.

dimensions.projects: one row per project.
dimensions.apps: one row per app, meaning one deployed version of a protocol that a project runs, such as uniswap-v3 or hyperliquid-spot.

Tables

Projects
Apps

dimensions.projects contains one row per project: the protocol or company behind one or more live apps.

Column	Type	Description
`project_id`	`STRING`	Identifier of the project. Every table that mentions a project joins on this.
`name`	`STRING`	Display name.
`symbol`	`STRING`	Ticker of the project’s token, where one exists.
`market_sectors`	`ARRAY<STRING>`	Every market sector tag the project carries.
`primary_market_sector`	`STRING`	The first element of `market_sectors`. One value per row.

dimensions.apps holds one row per app, meaning one deployed version of a protocol that a project runs, such as uniswap-v3 or hyperliquid-spot. A project can run several apps, and an app’s numbers add up into its project’s.

Column	Type	Description
`app_id`	`STRING`	Identifier of the app. Every table that mentions an app joins on this.
`project_id`	`STRING`	Project that runs the app.
`name`	`STRING`	Display name.
`market_sectors`	`ARRAY<STRING>`	Every market sector tag the app carries.
`created_at`	`TIMESTAMP`	When the app was added to the registry.

Sample queries

Project identity
Projects by sector
Apps for one project

Look up one project’s identity and sectors.

select
    project_id,
    name,
    symbol,
    primary_market_sector,
    market_sectors
from `dimensions.projects`
where project_id = 'uniswap'

Count projects per primary market sector.

select
    primary_market_sector,
    count(*) as projects
from `dimensions.projects`
group by primary_market_sector
order by projects desc

List every app one project runs.

select
    app_id,
    name,
    market_sectors,
    created_at
from `dimensions.apps`
where project_id = 'hyperliquid'
order by app_id

Notes

market_sectors is a list, so joining through unnest(market_sectors) yields one row per tag. A daily table joined that way counts a project once for every tag it carries, and any total you take off it comes out too high. The hazard is built into the shape of the query, whatever the tagging holds. Grouping by primary_market_sector keeps one row per project however the tagging changes later. Metrics shows the sector totals this protects.

Overview

Core entities

Market verticals

Chain verticals

Tables

Sample queries

Notes

​Tables

​Sample queries

​Notes

Tables

Sample queries

Notes