GA4Dataform Community: raw GA4 export into clean session and user models

Every analyst who works with GA4's raw BigQuery export ends up writing the same SQLX sooner or later: sessionise the event stream, resolve users across device and login boundaries, tidy purchase and engagement events into something a stakeholder can actually query. GA4Dataform Community is the fresh answer to that repeated chore, an open-source Dataform package that generates the session, user and event models for you instead of hand-rolling them per client.

Why it made the cut

Rather than shipping fixed SQLX files, the package generates its session, user and event Dataform models through a JS class layer, so the modelling logic is configurable rather than copy-pasted and re-edited for every new account. It is built and maintained by a named team of GA4 analytics engineers at Superform Labs, Jules Stuifbergen, Krisztián Korpa, Artem Korneev, Simon Breton and Johan van de Werken, and it is the acknowledged successor to the older dataform-ga4-sessions project, which is now deprecated. It is GPL-3.0 licensed, free and open source.

How to use it

1. Confirm the target GA4 property has the BigQuery export enabled and events are landing in events_* tables.
2. Set up Dataform on BigQuery for the project, if it is not already running.
3. Pull the package from the GitHub repository below and follow its configuration to point the generated session, user and event models at your export dataset.
4. Run the Dataform pipeline and review the generated models before building dashboards or further transforms on top of them.

Worth knowing

The current release is version 1.19.0, last updated 15 January 2026, with 73 commits and 84 stars on GitHub at time of writing. It replaces the older, deprecated dataform-ga4-sessions project, so if you inherit a client account still running that one, this is the migration target. As with any GPL-3.0 dependency, check the licence terms fit your delivery model before shipping it inside a client-facing product.