All data is aggregated and de-identified. Individual PatternLens conversations are never shown. Sensitive breakdowns are suppressed when cohort size is below the minimum threshold.