Public datasets, packaged for direct download.
Data Point is a lightweight catalog of open and public-domain datasets. No sign-up, no paywall β just clean files in the format you need.
Built for people who just want the file
Every dataset is normalized, documented, and versioned. Grab the whole thing or a single slice β no accounts required.
Direct downloads
Stable URLs backed by a global mirror network. Resume-friendly and scriptable with curl or wget.
Consistent schemas
Column names, types, and units are standardized across the catalog so joins just work.
Multiple formats
Each release ships as CSV, JSON, Parquet, and SQLite from a single source of truth.
Versioned releases
Pin to a snapshot or always pull the latest. Changelogs document every revision.
Clear licensing
Only open and public-domain sources, each with its license and provenance attached.
Searchable metadata
Rich tags, descriptions, and record counts help you find the right set fast.
Featured datasets
A sample of what's available right now. The full catalog lists every collection with its schema and history.
One source, four exports
Pick the format that fits your pipeline. All exports share identical columns and row counts.
CSV
UTF-8, RFC 4180 compliant, with a documented header row. Opens anywhere.
JSON
Newline-delimited (NDJSON) for streaming, plus a compact array variant.
Parquet
Columnar and compressed for analytics workloads and fast scans.
SQLite
A single-file database with indexes, ready to query offline.
Start downloading in seconds
Browse the catalog, copy a dataset URL, and pull it straight into your workflow. No keys, no quotas for public collections.
Open the catalog