Skip to content

Group 2 - Advanced I

Your slice of MediaPulse

The project you have seen before now splits in two. mediapulse_analytics is a second dbt project that doesn't own a staging layer of its own for news, podcasts, or ads - it reaches into mediapulse_base for that data via dbt Mesh using a cross-project dependency declared in dependencies.yml.

Your team owns the Streaming and Podcasts section of MediaPulse. You will work within that area to develop your project and your dbt skills.

Learning objectives

By the end of today you will be able to:

  • Use dbt Catalog's multi-project view to diagnose a bug that spans a project boundary, without reading the SQL first
  • Write singular tests and configure severity/thresholds for situations that shouldn't hard-fail a build
  • Explain what a custom generate_schema_name macro controls, and implement equivalent logic in a second project
  • Explain the difference between a package dependency and a project dependency, and what access: public actually permits
  • Run and interpret dbt-project-evaluator output across a single-project and a mesh-consuming project
  • Explain where Snowflake's dynamic data masking and dbt's role in applying it begin and end
  • Design a CI/CD approach (Slim CI, deferral) that accounts for a producer/consumer project relationship

Relevant projects

You'll work across both:

  • mediapulse_base - the producer project; note its access settings in dbt_project.yml and its stg_news__authors.email column
  • mediapulse_analytics - the consumer project; its dependencies.yml, and its fct_content_performance/fct_ad_revenue marts that ref() across the mesh boundary

See the MediaPulse overview for the underlying business context, and each project's own README.md for how they're structured and how they relate to each other.