# Data schema

Every `figures/figure-N.csv` and `data/aggregates.csv` in this report carries the
same ten columns, in this order, with one row per published value.

| Column | Meaning |
| --- | --- |
| `row_id` | A stable identifier for the row. In `data/aggregates.csv` it is prefixed with the figure it came from. |
| `label` | The row in plain words, as the figure and the table print it. |
| `value` | The number itself, with no unit attached. |
| `display` | The number as the figure, the table, and the Markdown twin print it, unit included. |
| `unit` | `percent` or `count`. |
| `numerator` | The count behind a rate. Empty when the value is not a rate. |
| `denominator` | The base the rate is taken over. Empty when the value is not a rate. |
| `ci_low` | The lower bound of the interval, where one was computed. |
| `ci_high` | The upper bound of the interval, where one was computed. |
| `suppressed` | `true` when the cell is under the reporting floor and the value is withheld. |

An empty cell means the value is missing or does not apply. A suppressed cell
carries `true` in the last column and no value.