Core Concepts¶

Understanding Dango's architecture, data flow, and key components.

Overview¶

Dango integrates four production-grade tools into a single platform:

• dlt - Data ingestion from 30+ sources (with access to 60+ via dlt_native)
• dbt - SQL transformations and data modeling
• DuckDB - Embedded analytics database
• Metabase - Business intelligence and dashboards

This section explains how these components work together to create a complete data platform.

What You'll Learn¶

Key Concepts at a Glance¶

Data Flow¶

Sources → dlt → DuckDB (raw)
                    ↓
              dbt transforms
                    ↓
             DuckDB (staging/marts)
                    ↓
               Metabase

All data layers live in DuckDB. dbt reads from and writes to DuckDB, transforming raw data into analytics-ready tables.

Three-Layer Data Architecture¶

1. Raw Layer - Immutable source data as-is
2. Staging Layer - Cleaned and deduplicated data
3. Marts Layer - Business metrics and analytics

Unified Interface¶

All functionality accessible through: - CLI - dango sync, dango start, etc. - Web UI - Monitoring and management at localhost:8800 - Direct SQL - Query DuckDB directly or use Metabase

Design Philosophy¶

Dango is built on two core principles:

Opinionated but Modular¶

Best practices are built-in so you can focus on insights, not infrastructure. As the open-source data ecosystem evolves, components can be swapped for better alternatives without rebuilding your entire stack.

Democratize Analytics Infrastructure¶

Enterprise-grade data tooling shouldn't require a dedicated platform team. Dango brings production-quality patterns to teams of any size.

Next Steps¶

Start with Architecture to understand how Dango's components work together, then explore the other pages to dive deeper into specific aspects.