# BGraph synthetic AI search signals

This package demonstrates aggregated AI discovery signals across 12 synthetic category and market combinations. It uses broad platform classes, query topics, intent classes, destination domains, cohort measures, and downstream visit metrics.

## Privacy boundary

- No raw prompt text.
- No private conversation text.
- No person-level profile or direct identifier.
- Query topics and intent classes are aggregated synthetic labels.

## Synthetic category examples

consumer electronics, travel, productivity software, smart home, financial services, apparel, beauty and personal care, online education, automotive, media streaming, cybersecurity, and food delivery.

## Limitations

- Every count, rate, category, market, and domain is fabricated.
- The sample does not claim current coverage, users, panel size, traffic, history, or product availability.
- Production cohorts require documented thresholds, suppression, source rights, jurisdiction, and intended-use review.
- Sensitive topics and small cohorts must be excluded or generalized.

## Files

- sample.csv
- sample.json
- parquet-schema.json
- data-dictionary.json
- example-query.sql