Data Build Tool - a tool to build data

The guest of our show presented the Data Build Tool, DBT, in one of our knowledge-sharing sessions. Especially the fun part is what I recognize when I look at this knowledge-sharing session. The guest of the show presented it in an enthusiastic way and used terms like: the macro functionality makes DBT awesome and Ephemeral is the Unique Selling Point.

What this episode covers

In General

• What is a Data Build Tool?
• What problems does it solve?
• What are the two versions? Cloud, CLI
• Why is it much more interesting to use than other ETL, transformation tools?
• Analytics Engineering
• Explain a bit more in debt the Basics of a DBT project: dbt project data pipeline, data model, tests, macros.
  • Model dependencies
  • DBT-packages ecosystem
• One of the powers of DBT is Jinja.
  • What could you compare Jinja to?
  • Is it easy to comprehend and learn?
• Ephemeral (CTE) was the unique selling point for you, why?
• An engineer enthusiastic about a documentation feature… that asks for an explanation.

Within bol.com

• Usage started as an experiment, how are we using it in bol.com?
• Can you share some learnings you and your team had when you started using DBT?

Guests

Sander Boumeester – Software Engineer in the experimentation team.

Notes

On the website of getdbt there's a cool picture that explains the framework really well.

Episode source