Lightdash

Hamzah Chaudhary and Oliver Laslett built Lightdash after watching data teams struggle with metrics drifting between transformation scripts and visualization layers. They designed the platform to turn dbt repository definitions directly into an analytics layer, keeping dimensional logic in version control. Today, that origin means engineering teams manage business intelligence dashboards through Git commits rather than rebuilding metric definitions across multiple proprietary interfaces.

Where Lightdash fits

Proprietary platforms like Looker and Tableau decouple visualization from the data engineering workflow, requiring proprietary modeling like LookML or complex manual dashboard configurations. Lightdash anchors directly to dbt, turning warehouse transformation models into a centralized semantic layer. Analysts define dimensions, metrics, and joins in dbt YAML files, querying warehouses like Snowflake, BigQuery, Redshift, and PostgreSQL without duplicate ingestion layers.

The trade-off is architectural dependency. Lightdash provides no spreadsheet-style data blending or flat file uploads; your warehouse must store the data and your dbt repo must model it. In exchange, it integrates with GitHub and GitLab to generate preview environments on every pull request, allowing developers to catch breaking schema changes before dashboards hit production.

What it does well

• Git-backed semantic layer: Dimensions, metrics, and joins live alongside dbt models in YAML, letting data teams review analytical logic via standard pull requests.
• Ephemeral preview instances: GitHub and GitLab actions spin up temporary Lightdash environments on branch updates, letting agencies test client dashboards before merging changes.
• Automated metric lineage: Visualizations inherit descriptions and tests from upstream dbt models, providing end users with data freshness context without manual upkeep.
• Scheduled channel delivery: Automates recurring chart and dashboard exports directly to Slack channels or email distribution lists.
• Multi-tenant space permissions: Role-based access controls and partitioned spaces enable agencies to isolate client dashboards within a single deployment.
• Controlled self-service querying: Non-technical users explore, filter, and drill into underlying warehouse rows within boundaries defined by pre-approved metrics.