# Data Dictionary

These fields appear in `outbreak_signals.csv` and in the `properties` object of each feature in `outbreak_signals.geojson`.

| Field | Description |
| --- | --- |
| `id` | Stable event identifier from the reviewed signal feed. |
| `country` | Country or territory associated with the signal. |
| `country_code` | ISO-style three-letter country code when available. |
| `region` | Subnational region when the source supports it. |
| `city` | City or locality when the source supports it. |
| `latitude` | Decimal latitude used for map display. |
| `longitude` | Decimal longitude used for map display. |
| `geometry_level` | Geographic precision, such as `country`, `region`, or `city`. |
| `signal_kind` | Internal category for the signal, such as public response, monitoring, or case-related context. |
| `signal_label` | Human-readable signal label shown or summarized on the map. |
| `evidence_level` | Evidence label: `official`, `verified`, `reported`, or `context`. |
| `mention_count` | Number of reviewed mentions grouped into this map event. |
| `source_count` | Count of distinct source domains or source groups for the event. |
| `official_source_count` | Number of official-source domains or source groups associated with the event. |
| `news_source_count` | Number of non-official public reporting source domains or source groups associated with the event. |
| `primary_source_type` | Primary source class used for summary display, such as `official`, `news`, or `unknown`. |
| `latest_published_at` | Latest source publication date in the grouped event. |
| `latest_checked_at` | Latest time the signal feed checked or exported the grouped event. |
| `first_seen` | Earliest date this grouped signal was seen in the review window. |
| `last_seen` | Latest date this grouped signal was seen in the review window. |
| `case_count` | Case count only when the grouped event supports one. Blank values should not be treated as zero. |
| `death_count` | Death count only when the grouped event supports one. Blank values should not be treated as zero. |
| `display_strength` | Display confidence used by the map, such as `solid`, `medium`, or `faint`. |
| `source_domains` | Semicolon-separated source domains associated with the grouped event. |
| `top_source_names` | Semicolon-separated names for up to five representative sources. |
| `top_source_urls` | Semicolon-separated URLs for up to five representative sources. |
| `top_source_titles` | Semicolon-separated titles for up to five representative sources. |

## Evidence Labels

- `official` - Directly supported by an official public health source or agency statement.
- `verified` - Supported by multiple credible sources or strong public health attribution.
- `reported` - Relevant public report that should be checked before being treated as confirmed transmission.
- `context` - Supporting response, monitoring, travel, healthcare, or background context.

## Important Interpretation Notes

Blank `case_count` and `death_count` fields do not mean zero. They mean the grouped signal should not be converted into a numeric case or death total.

Country-level coordinates are map anchors for grouped public reports. They should not be interpreted as exact exposure sites, patient homes, treatment facilities, or confirmed local transmission locations.