# Data dictionary

All revenue fields are in EUR and are gross royalties through the data cut (2026-04-19), not profit. See `METHODOLOGY.md` for definitions and limitations.

## `catalog_books.csv` (540 rows)

| Field | Type | Description |
|---|---|---|
| `book_id` | string | Anonymous stable ID, `BK0001`–`BK0540`. No title or ASIN. |
| `pen_id` | string | Anonymous pen-name cluster, `PN01`–`PN08`. Unicode variants merged. |
| `primary_pen` | boolean | `true` for the pen name with the largest catalog (441 titles). |
| `era` | enum | `ai_experiment` (published 2025-03 or later) or `legacy` (before 2025-03). Temporal label, not a per-book AI provenance flag. |
| `cohort` | string | Publication month, `YYYY-MM`. |
| `status` | enum | `winner`, `active`, `watch`, `dead`, `too_new`. See thresholds below. |
| `lifetime_eur` | number | Lifetime royalties in EUR through 2026-04-19, including the KENP estimate. |
| `last_90d_eur` | number | Royalties in the 90 days before 2026-04-19. |
| `revenue_band` | enum | Convenience band for `lifetime_eur`: `0`, `0-10`, `10-50`, `50-100`, `100-500`, `500+`. |

### Status thresholds

| Status | Rule |
|---|---|
| `winner` | last 90 days > €30 |
| `active` | last 90 days €10–30 |
| `watch` | last 90 days €2–10 |
| `dead` | last 90 days < €2 |
| `too_new` | published after January 2026 |

The label is derived from the last-90-day snapshot, not from lifetime revenue or quality.

## `monthly.csv` (16 rows)

| Field | Type | Description |
|---|---|---|
| `month` | string | `YYYY-MM`, 2025-01 to 2026-04. |
| `total_eur` | number | Total royalties for all catalog titles in the month. Includes legacy titles. |
| `active_books` | integer | Titles with recorded activity in the month. |
| `catalog_size` | integer | Titles live in the catalog at that point. |
| `eur_per_catalog_book` | number | `total_eur / catalog_size`. A blunt efficiency measure; it does not account for titles that are too new to sell. |

The monthly table starts in 2025-01. Pre-2025 legacy revenue (€5,307) is in the lifetime totals but not in this file.

## `cohorts.csv` (23 rows)

| Field | Type | Description |
|---|---|---|
| `cohort` | string | Publication month. |
| `era` | enum | `ai_experiment` or `legacy`. |
| `books_published` | integer | Titles published in the cohort. |
| `total_lifetime_eur` | number | Sum of lifetime revenue for the cohort. |
| `median_lifetime_eur` | number | Median lifetime revenue for the cohort. |
| `median_first_month_eur` | number or blank | Median revenue in the first full calendar month. Available for AI-era cohorts; blank for legacy cohorts. |
| `pct_above_50_lifetime` | number | Share of the cohort with lifetime revenue ≥ €50, in percent. |
| `pct_above_20_lifetime` | number | Share of the cohort with lifetime revenue ≥ €20, in percent. |

## `pennames.csv` (8 rows)

| Field | Type | Description |
|---|---|---|
| `pen_id` | string | Anonymous pen-name cluster. Names are not included. |
| `books` | integer | Titles under the cluster. |
| `lifetime_eur` | number | Lifetime royalties for the cluster. |
| `last_90d_eur` | number | Last-90-day royalties for the cluster. |
| `median_lifetime_eur` | number | Median lifetime revenue across the cluster's titles. |
| `winners` | integer | Titles labelled `winner`. |
| `dead` | integer | Titles labelled `dead`. |
| `eur_per_book_lifetime` | number | Lifetime revenue divided by titles. |
| `eur_per_book_last_90d` | number | Last-90-day revenue divided by titles. |
| `share_lifetime_pct` | number | Share of total lifetime revenue, in percent. |

## `status.csv` (5 rows)

| Field | Type | Description |
|---|---|---|
| `status` | enum | `winner`, `active`, `watch`, `dead`, `too_new`. |
| `books` | integer | Number of titles with the label. |
| `share_pct` | number | Share of the 540 titles, in percent. |
| `definition` | string | Human-readable threshold. |
| `min_last_90d_eur` | number | Minimum last-90-day revenue in the group. |
| `max_last_90d_eur` | number | Maximum last-90-day revenue in the group. |

## `revenue_bands.csv` (6 rows)

| Field | Type | Description |
|---|---|---|
| `band` | enum | Lifetime revenue band. |
| `books` | integer | Titles in the band. |
| `share_pct` | number | Share of the 540 titles. |
| `lifetime_eur` | number | Total lifetime revenue in the band. |
| `share_lifetime_pct` | number | Share of total lifetime revenue. |

## `summary.json`

Recomputed headline metrics, including:

- `totals`: books, era split, lifetime/last-90-day revenue, mean and median, share under €10 and at least €100;
- `status`: counts and shares;
- `concentration_last_90d_pct`: top 10, 25, 50, 100 and 200 titles' share of last-90-day revenue;
- `monthly`: month count, total, peak month, pre-2025 legacy revenue, selected per-book values;
- `cohorts`: cohort count, July 2025 hit rate, November 2025+ weighted hit rate;
- `pennames`: count and largest cluster;
- `revenue_bands`;
- `corrections`.

## `datapackage.json`

[Frictionless Data](https://datapackage.org/) package metadata: name, title, description, version, license, sources, contributors, resource paths, row counts and the `catalog_books.csv` field schema.

## Missing values

- `cohorts.csv` `median_first_month_eur` is blank for legacy cohorts because the source workbook's first-month comparison begins with the 2025 cohorts.
- No other public field is blank. A zero means zero recorded revenue at the cut, not missing data.
