PROVIDER - Metrics as code BI for dbt
Lightdash
Lightdash puts metric definitions back into dbt: dimensions and metrics live in versioned models and drive dashboards directly, removing a second definition set. With self hosting or cloud, it fits teams already on dbt, while non-dbt teams and complex visualisation needs gain less.
3.8pts
Detailed Notes
Pros 5 items
Cons 4 items
Service Overview
Lightdash is a BI product from a United Kingdom team, representing the modelling as code branch of data visualisation and BI. Metrics are not stored inside the BI tool; they live in dbt models, and the BI layer only presents that same definition.
The test is simple: is the team already running dbt. If yes, Lightdash removes the double maintenance of one definition set in dbt and another in the BI tool. If no, adopting dbt is the real prerequisite.
Key Strengths
- One source of truth: metric definitions share the repository and version with dbt
- Flexible deployment: self hosted or managed cloud
- Generated dimensions: fields and descriptions come from dbt metadata
- Permissions and portal: row level control plus a content portal
- Auditable code: the open source edition can be reviewed for compliance
Product Ecosystem
Platform capabilities
dbt project connection, metric and dimension definitions, dashboards and charts, a content portal, row level permissions, versioning and query caching.
Integrations
Works with dbt Cloud or dbt Core, common warehouses and CI flow, so metric changes pass through code review with the model.
Typical scenarios
Data teams already on dbt, metric boards that demand strict consistency, and organisations with self hosting compliance requirements.
Limitations
- dbt dependency: teams without it gain little
- Restrained visuals: complex charts need another tool
- Operations cost: self hosting means running the service
- Ecosystem size: fewer connectors than mature commercial BI
Use Cases
- Teams already on dbt (★★★★★): the shortest route to one definition
- Self hosting requirements (★★★★☆): compliance friendly
- Complex visualisation (★★☆☆☆): pair with a charting tool
- Pure business self-service (★★★☆☆): dbt literacy still needed
Pricing
| Item | Billing model | Notes |
|---|---|---|
| Cloud service | Per seat subscription | Hosting and permissions included |
| Self hosted | Open source, no licence fee | Operations are yours |
| Enterprise | Contact sales | Governance and support |
FAQ
Is dbt required?
Yes, Lightdash reads metric definitions from dbt models.
Can it be self hosted?
Both self hosted and managed cloud are available.
How do we see metric change impact?
Definitions follow code versions, so they are reviewed and traced like code.
Are chart types sufficient?
Common charts are covered; deep customisation belongs with another tool.
Code Examples
# Define metrics in a dbt model; Lightdash reads this definition directly
models:
- name: orders
meta:
metrics:
revenue:
type: sum
sql: amount
description: Total paid order amount
columns:
- name: campaign
meta:
dimension:
type: string
# Verify the dbt project metadata parses correctly
dbt parse --project-dir ./analytics --profiles-dir ~/.dbt
Community & Resources
- Website: https://www.lightdash.com/
- Documentation: https://docs.lightdash.com/
- Source repository: https://github.com/lightdash/lightdash