Week 10: dbt Transformations

Welcome to Week 10! This week introduces dbt (data build tool), the industry standard for SQL-based data transformations. You will set up dbt for Azure PostgreSQL, learn how Jinja templating extends SQL, write tests to ensure data quality, and build your first mart that will power next week's dashboard.

You continue working with the NYC taxi dataset from Week 9 (raw_trips and raw_zones already loaded in Azure PostgreSQL).

Learning goals

Install and configure dbt Core to work with Azure PostgreSQL
Understand the dbt project structure: models, sources, profiles, and macros
Write SQL models enhanced with Jinja templating for DRY, reusable transformations
Explain materializations (view, table) and when to use each; understand incremental as a forward reference
Build a staging/intermediate/mart layer structure with a mart that powers a dashboard
Implement data tests (not_null, unique, relationships) and unit tests for transformation logic
Generate and navigate dbt documentation to understand data lineage

Prerequisites

Completed Week 9 with raw_trips and raw_zones loaded into the shared Azure PostgreSQL instance
PG_PASSWORD environment variable and access to your personal dev_<your_name> schema
uv installed (used to install dbt-core + dbt-postgres)

Chapters

Supplementary

Career relevance: Week 10 in the NL data job market

Glossary

History of Analytics Engineering

Going Further: Optional Deep Dives