Lightdash

Every BI tool asks you to define metrics inside it. If you already define them in dbt, you now maintain two definitions of revenue, and eventually they disagree — usually in a board meeting. Lightdash removes that by reading metrics directly from the dbt project.

It is a British company, MIT-licensed and fully self-hostable, which makes it the natural BI layer for a European data stack built on dbt and a warehouse of your choice.

Definitions in Version Control

Metrics and dimensions are declared in dbt YAML alongside the models, reviewed in pull requests and versioned with everything else. Lightdash exposes them for exploration, so analysts and business users self-serve against definitions that engineers control.

That governance model — one definition, reviewed, versioned — is what Looker popularised, and Lightdash provides it without Looker's cost or Google's ownership.

Self-Hosting

A Docker deployment on your own European infrastructure keeps the whole analytics layer under your control. Combined with a warehouse in an EU region, no analytical data or metric definition needs to leave European jurisdiction at any point.

Practical Features

Charts, dashboards, scheduled delivery, Slack alerts on metric thresholds, and an API for embedding. The exploration interface is aimed at business users who should not have to write SQL, while analysts can drop into SQL when needed.

Limitations to Consider

Lightdash is only sensible if you use dbt — without it, the central advantage disappears and Metabase or Cluvio will be a better fit. It is a young product with a smaller feature surface than Tableau or Power BI, particularly for complex visualisation and pixel-perfect reporting. Self-hosting means you operate it.

Who Should Use Lightdash

Data teams with a modern stack — warehouse plus dbt — who want self-service BI with governed definitions, and European organisations that want the analytics layer self-hosted rather than with a US vendor.