dataset_bundle/schema_notes.md

# Schema Notes

## events.csv
- `event_date`: observed date when extractable, otherwise a publication-date proxy.
- `review_status`: public rows are auto-included deterministic event rows only.
- `source_count`: number of supporting stories retained in the dedupe cluster.
- `language_family`: deterministic source-language bucket such as official UAS report, ordinary drone report, mystery-drone language, or explicit UFO/UAP language.
- `uap_overlap_label`: conservative language-overlap label; it is not a verification that the object was a UAP.
- `ambiguity_score`: deterministic language score, not a confidence score or factual finding.

## monitored_areas.csv

- Rows represent official covered-area points or centroids used for event plotting.



## event_sources.csv
- Lists supporting story URLs and titles retained for each public event.