Preply Data · Docs · Home

Preply Data Docs

Curated knowledge base for business-critical dbt models, methodologies and domain concepts — the why and how to use, on top of schema.yml and Open Metadata.

— docs — domains Updated — GitHub
01 · Overview

At a glance

Every page here is a hand-picked, human-readable one-pager for a dbt model or data concept that needs more than a column description. Simple staging models stay documented in schema.yml.

Total pages
— curated docs
Domains
— business areas
Ownership
Analytics Engineers
Publication
Auto on merge data-dbt repo · anything in docs/ on main
Complements, not replaces
This knowledge base sits on top of dbt schema.yml (column-level, exhaustive) and Open Metadata (lineage and governance catalogue). Use it for the why, the methodology, and the non-obvious caveats.
02 · Browse

Find a doc

⌕
03 · Contribute

Add or update a doc

Docs are generated via the shared Cursor skill dbt-model-html-docs. Place the file in the correct domain folder (docs/<Domain>/<model_name>.html) and open a PR — the page is published automatically on merge to main.

  1. Decide it's worth a page. CDS tables backing multiple dashboards, non-trivial methodologies (projections, allocations, attribution, SCD2), or tables with historical caveats that don't fit in a column description.
  2. Run the skill in Cursor. Invoke dbt-model-html-docs; it walks the SQL, schema.yml, lineage, and macros, asks you for external links, and fills the template.
  3. Open a PR. Drop the generated HTML under the right domain folder. Match existing folder names — don't silently introduce a new top-level domain.
  4. Merge and share. On merge, the page is published. Post the link in the owning domain's Slack channel and, if relevant, cross-link from the model's schema.yml description.
Quality bar
No fabricated URLs, table names, or PR numbers — every claim must trace to SQL, schema.yml, a macro, or a link explicitly provided by the owner. Consistency beats creativity: keep the structure and styling from the skill.
04 · Resources

Where else to look

The three technical-metadata layers at Preply Data, and how this site fits between them.

Source of truth
dbt repo · schema.yml
Column descriptions, tests, contracts and lineage for every model.
Catalogue
Open Metadata
Tables, columns, owners, governance and cross-warehouse lineage.
Consumption
Looker
Explores and dashboards that surface these models to stakeholders.
Curation skill
dbt-model-html-docs
The Cursor skill that generates pages on this site. Lives at .cursor/shared_skills/dbt-model-html-docs/.
Source code
data-dbt · /docs ↗
The GitHub folder where every page on this site lives. Open a PR to add or update a doc.