Glossary
BigQuery public dataset
A BigQuery public dataset is a table collection Google hosts for anyone to query with SQL, offered as an alternative to downloading files. Examples include the patents.publications table at 98,176,830 rows (about 899 GB) and PyPI's Linehaul download logs. The catalog shows several shapes of this.
What is BigQuery public dataset?
A BigQuery public dataset places a large structured table inside Google's warehouse program so analysts query it in place rather than mirroring files. The catalog shows several shapes of this. Google Patents Public Data documents its patents.publications table at 98,176,830 rows and about 899 GB, with exports to CSV, JSON, Avro and Parquet via BigQuery export and even Google Sheets via connected sheets; querying is free within BigQuery's monthly free tier of 1 TB processed. PyPI streams download logs from Linehaul into BigQuery tables spanning 877,215 projects and 44.8 TB of package size, cataloged by Datadory. OSV lists BigQuery among its data formats, GitHut was built on the githubarchive public datasets, and the Google Ads Transparency Center notes its BigQuery public datasets fall under Google Cloud Terms of Service.
Why does BigQuery public dataset matter when choosing a dataset?
Hosting shifts the economics: storage becomes someone else's problem, while every byte queried draws down quota. At patents scale - 98 million rows and 899 GB - a careless SELECT over wide columns can consume meaningful fractions of the 1 TB monthly free tier, so cost discipline becomes part of data work. Terms also follow the platform: the Ads Transparency Center explicitly subjects its BigQuery public datasets to Google Cloud Terms of Service, layered on whatever licence governs the content (PyPI's tables carry Creative Commons terms). The upside is real - SQL joins across enormous corpora plus native export paths - but buyers should budget queries, not just downloads, and keep a file export path in mind.
How do you evaluate BigQuery public dataset in a data source?
- Check documented table size first: patents.publications is documented at 98,176,830 rows / about 899 GB - sizes drive query cost more than row counts suggest.
- Check the free-tier math: querying is free within BigQuery's 1 TB monthly processing tier; anything larger needs billing attached.
- Check the governing terms: the Google Ads Transparency Center states its BigQuery public datasets are subject to Google Cloud Terms of Service.
- Check export routes off the platform: CSV / JSON / Avro / Parquet via BigQuery export, plus Google Sheets via connected sheets, keep you portable.
- Check licence provenance: PyPI's tables are cataloged by Datadory.
Related terms
Related entries on bulk access paths:
Frequently asked questions
What is a BigQuery public dataset?
A table set Google hosts for SQL access under its public-dataset program. Examples include patents.publications at 98,176,830 rows (~899 GB) and PyPI's Linehaul download-log tables covering 877,215 projects.
Does querying a BigQuery public dataset cost money?
Within BigQuery's monthly free tier of 1 TB processed, no; beyond it, normal query charges apply. Large tables like the 899 GB patent corpus make query design matter for cost.
Can I export data out of BigQuery?
Yes. Formats include CSV, JSON, Avro and Parquet via BigQuery export, and Google Sheets through connected sheets, so analysis need not stay inside the warehouse.
Datasets containing this field
Datasets containing BigQuery public dataset
6 datasets carry bigquery public dataset in the catalog. Open one, count the fields, judge for yourself.
Data.gov Catalog - Advertising Datasets
FEC Campaign Finance Data & OpenFEC API
Google Ads Transparency Center
Hugging Face Datasets - Advertising Search
Every listing shows the field dictionary, sample rows, and coverage before you commit. API, files, or your warehouse. Daily, weekly, or hourly.
Get sample rows