The GA4 sample ecommerce dataset: a free BigQuery sandbox for practising SQL

Every GA4-BigQuery tutorial and interview take-home test reaches for the same table sooner or later, because most people learning this stack do not have a live GA4 property with the BigQuery export switched on, let alone months of real event history to query. Google's public sample dataset closes that gap: it is the standard practice ground for writing GA4 SQL before you ever touch a client account.

Why it made the cut

The dataset holds 92 days of obfuscated real GA4 event-export data from the Google Merchandise Store, spanning November 2020 to January 2021. It is Google's own evergreen, officially maintained practice dataset, not a third-party mirror, and it uses the genuine event_params nested/repeated schema you will meet in any production GA4 export. Query it with BigQuery Sandbox or a free-tier project, no billing account required for light use, and you have a working GA4 event stream to test joins, funnels and aggregations against.

How to use it

1. Open a Google Cloud project with BigQuery access, the free Sandbox tier is enough.
2. In the BigQuery console, add the public dataset bigquery-public-data.ga4_obfuscated_sample_ecommerce to your project (or query it directly by fully qualified name).
3. Run a SELECT against the daily events_* tables, filtering on event_date and unnesting event_params where you need a specific parameter.
4. Use it to rehearse the standard GA4 patterns: sessionisation, purchase funnels, e-commerce revenue rollups, before running the same SQL against a live property.

Worth knowing

This is the dataset behind the Vault's GA4 Data-Quality Auditor SQL: run that script against this sandbox first to see the expected output shape before pointing it at a real property. The Google docs page was last updated 29 June 2026, and the dataset itself has been the community's reference practice set since GA4's BigQuery export launched.