GA4 bronze catalog and warehouse modeling
Skipprd's GA4 source lands daily fact grains in bronze via the Data API runReport. Each namespace is one stable dimension set; rollups, pivots, and channel summaries belong in the warehouse (dbt or SQL).
This is not the GA4 BigQuery export (event-level raw). Use a separate future source for event-level hoovering.
Bronze → warehouse examples
| Bronze namespace | Example warehouse outputs |
|---|---|
google_analytics.audience_daily | Site-wide DAU, sessions, engagement KPIs |
google_analytics.audience_retention_daily | WAU / 28-day active users (metrics already on date grain) |
google_analytics.traffic_acquisition_daily | Channel summary, source/medium rollups |
google_analytics.traffic_campaign_daily | Paid campaign performance |
google_analytics.user_acquisition_daily | New user acquisition by channel |
google_analytics.events_daily | Weekly event trends (group by week in SQL) |
google_analytics.content_pages_daily | Top pages, landing page analysis |
google_analytics.geo_daily | Country / region / city dashboards |
google_analytics.demographics_*_daily | Age, gender, interest, language breakdowns |
google_analytics.ecommerce_items_daily | Product-level revenue (requires ecommerce enabled) |
google_analytics.publisher_ads_daily | Ad unit performance (requires linked Ads) |
Accuracy knobs (do not conflate)
| Setting | Role |
|---|---|
replace_partition on date | Rewrite each calendar day’s partition when metrics change |
lookback_days | Re-pull recent mature days Google may revise |
processing_lag_days | Skip syncing the trailing edge while GA4 is still incomplete |
window_in_days | Days per API dateRanges chunk; default 1 to limit sampling |
Recommended operations: scheduled skipprd sync with defaults; raise lookback_days for long attribution windows — not window_in_days > 1.
See Google Analytics (GA4) input and Source landing semantics.