# Tomás's Governance Dashboard — Schema

Flat CSV in `tomas-governance.csv`. One row per Page-1 visual.

## Columns

| Column | Type | Description |
| --- | --- | --- |
| `id` | integer | Visual id (1..15) in the order Tomás built the page. |
| `kind` | text | Chart family (`kpi_card`, `trend_line`, `treemap`, `heatmap`, `donut`, `map`, `bar_chart`, `table`). |
| `content` | text (quoted) | Short description of what the visual shows. |
| `verdict` | text | Either `keep` (survives the decluttering pass) or `remove` (does not). |
| `reason` | text (quoted) | The one-line justification for the verdict. |

The `content` and `reason` fields are CSV-safe quoted strings so commas inside descriptions do not break parsers.

## Row inventory

- 15 rows — one per visual on the original fifteen-visual page.

Total: 15 rows.

## Deliberate design choices

The dataset is structural rather than numeric. Its purpose is to lock the shape of the fifteen-to-seven decluttering pattern so the reader can reproduce Tomás’s verdict distribution and see the reasoning per visual. The `reason` column is where the pedagogical value lives — naming *why* each visual survives or is removed teaches the discipline better than the count itself.

Chart-family diversity is preserved on purpose: the kept set spans kpi_card, trend_line, heatmap, bar_chart, and table (five families across the five kept visuals). A rebuild that collapsed to a single family would teach the wrong lesson.

## License

Released for teaching use alongside *The Defensible Decision*. Use freely.