From Data Analyst to Analytics Engineer

What

You will build a small commerce pipeline that answers one question: "How much revenue did each customer generate each day?"

The project starts with local order and customer files. You will inspect the source, clean it in a staging layer, model orders and customers, build a daily revenue table, add checks, and make repeat runs safe. The final lessons cover documentation, Git review, local pipeline runs, a high-level look at automation, correcting a changed source record, and a capstone based on returns.

CSV source data
  -> staging models
  -> order fact + customer dimension
  -> customer daily revenue table
  -> checks, documentation, local runs, and automation options

By the end, you will have a working project plus the queries, checks, and review notes used to verify it.

How

Every lesson starts with an analytics engineering decision, then shows one way to implement it with Bruin and DuckDB. The ideas transfer to dbt, Dataform, SQLMesh, or an in-house framework: define what each row means, make sure joins do not duplicate rows, test assumptions, document metrics, review changes, and run the pipeline locally before deciding how to automate it.

Use the same project throughout the course and complete the checkpoint at the end of each lesson. Do not skip the small queries that compare source and output rows. They are the habit that turns a plausible model into a trustworthy one.

You can complete every exercise manually or ask a coding agent such as Codex, Claude Code, or Cursor to edit files and run the same commands. The manual instructions remain the source of truth. When you use an agent, review its proposed changes, inspect the diff, and read the command output before moving on.

Python assets and agent-assisted work are optional extensions. Finish the SQL workflow first, then use the additional resources if they match your job.