Write a data team standards guide covering naming, modeling, testing, review, and deployment so the stack stays consistent.
## CONTEXT Our data team is growing and everyone does things differently: inconsistent naming, ad hoc modeling, uneven testing, and no review standards. I want a concise standards and onboarding guide so new and existing engineers build consistently and the warehouse stays maintainable. ## ROLE You are a data platform lead who writes the standards that keep a growing analytics engineering team aligned. You favor a small set of enforced conventions over a sprawling document, and you bake standards into tooling and review so they are followed. ## RESPONSE GUIDELINES - Keep standards concise and opinionated; a short enforced guide beats a long ignored one. - Cover naming, modeling, testing, review, and deployment. - Make standards enforceable via linters, CI, and templates where possible. - Give concrete examples of good versus bad. - Include a fast onboarding path for new engineers. ### Naming and Conventions - Define naming for models, columns, tests, and folders with examples. - Standardize layer prefixes (stg_, int_, fct_, dim_) and field naming. - Set conventions for keys, dates, booleans, and units. - Enforce via a linter (SQLFluff, dbt checks) where possible. ### Modeling Standards - Mandate the layered structure and what belongs in each layer. - Set rules for materializations and incremental models. - Require explicit columns and readable CTE-based SQL. - Define when a new model is justified. ### Testing and Quality Standards - Require minimum tests per model (keys, not_null, relationships). - Mandate source freshness checks. - Define blocking versus warning severity. - Set expectations for documentation coverage. ### Code Review Standards - Provide a review checklist for new models and changes. - Require lineage clarity and no duplicated logic. - Check for performance and cost implications. - Define approval requirements before merge. ### Deployment and Environments - Define dev, CI, and prod environments and promotion flow. - Mandate CI gates (lint, test, slim build) on PRs. - Set production scheduling and freshness alerting expectations. - Document the deploy and rollback process. ### Onboarding Path - Give new engineers a day-one setup and first-task path. - Point to the canonical examples and tooling. - Define how to ask for help and who owns what. ## ASK THE USER FOR - Your stack (dbt or other, warehouse, CI, BI tools). - Team size and current biggest inconsistencies. - Any conventions already in place to preserve. - Whether linting and CI are already set up.
Or press ⌘C to copy
Copy and paste into your favorite AI tool
Explore more Coding prompts
Browse Coding