Introduction to dbt

dbt (data build tool) lets analysts write data transformation logic as SQL SELECT statements, then handles materializing those queries as tables or views in your warehouse, running tests, and generating documentation — all from the command line or dbt Cloud.

Placeholder Lesson

This lesson is a work in progress. Content will be expanded with a full project walkthrough, exercises, and a video.

Where dbt Fits

The modern data stack typically looks like this:

Raw data sources
      ↓
  Ingestion (Fivetran, Airbyte)
      ↓
  Data Warehouse (Snowflake, BigQuery, Redshift)
      ↓
  Transformation → dbt ← you are here
      ↓
  BI / Analytics (Metabase, Looker, Tableau)

dbt operates entirely inside your warehouse. It never moves data — it only transforms data that is already there.

Key Concepts

Concept	What it means
Model	A .sql file containing a single SELECT statement
Materialization	How dbt persists a model: view, table, incremental, or ephemeral
ref()	How one model references another — dbt builds the DAG automatically
Test	An assertion about your data (e.g., no nulls, unique keys)
Source	A raw table in your warehouse declared in sources.yml

Your First dbt Model

Create a file at models/staging/stg_orders.sql:

-- models/staging/stg_orders.sql
select
    order_id,
    customer_id,
    order_date,
    total_amount
from {{ source('raw', 'orders') }}

Run it with:

dbt run --select stg_orders

What's Next

In the next lesson you'll learn how to structure a dbt project with staging, intermediate, and mart layers.