Skip to main content

Data Model Documenter

Exports a Tableau data source's logical model to a shareable markdown file. Read-only plus one file write. Current-state snapshot; no diff.

Example: You point it at a Snowflake data source with five tables and 12 relationships. It writes snowflake_prod_model.md next to the workbook with a table inventory (name, row-count estimate, field count), relationship map (cardinality, join conditions), calc-dependency graph, and a per-table field appendix. You share the file with a coworker who doesn't have Tableau access.

What it does​

The logical model lives inside the .twb and is hard to share. Documenter reads workbook state, extracts the model, and produces a single shareable markdown output. Deterministic; two runs on the same source produce the same file (modulo row-count sample variance on live sources).

When to use it​

  • "Document this data model."
  • "Export the tables and relationships."
  • "Share this data source structure."
  • "Make a data model doc for handoff."

When NOT to use it​

  • Field-meaning documentation (use data-scribe).
  • Change tracking over time (use workbook-auditor).
  • Recommending model changes (use join-and-relationship-advisor).
  • Connection setup (use data-source-connector).
  • Data-content audit (use data-quality-review).

What you get back​

  • Markdown file written to disk (default: {datasource-name}_model.md next to the workbook).
  • Sections: data source metadata, table inventory (with row-count estimates), relationship map (cardinality + join conditions), calc-dependency graph, per-table field appendix.
  • Chat summary: file path plus one-line inventory of what was documented.

Sections in the Output File​

  1. Data source metadata: name, connection type, extract vs. live, extract config summary.
  2. Table inventory: each table with row-count estimate and field count.
  3. Relationship map: each pair with cardinality, join condition, referential-integrity assumption.
  4. Calc-dependency graph: each calc with formula and referenced fields.
  5. Appendix: per-table field list with datatype and role.

Limitations​

  • Row-count estimates on live connections may not equal true source count.
  • Cannot detect model changes over time (workbook-auditor).
  • Cannot document field meaning beyond table-and-datatype level (data-scribe).
  • Cannot document sets, hierarchies, or parameters that don't surface via list-available-fields. Noted as "not extractable."
  • One data source per run.
  • No publishing anywhere. Desktop file write only.