datapond registry

Central registry of curated DuckDB databases built from public government and research data.

What is datapond?

datapond is a collection of clean, queryable DuckDB databases built from messy public data sources. Each database is:

• Reproducible -- built from public source files with a scripted pipeline
• Queryable -- stored as a single .duckdb file with documented tables
• Accessible -- every database can be queried remotely in seconds with no download required, or downloaded locally for full speed
• Documented -- includes a _metadata table and full README

Quick start

pip install datapond

import datapond

# See what's available
datapond.list()

# Connect and query instantly (streams over HTTP, no download)
con = datapond.connect('eoir')
con.sql("SELECT * FROM proceedings LIMIT 5").show()

# Download for local use
datapond.download('eoir')

Available databases

Database	Rows	Tables	Size	Source
eoir	164.6M	98	6.6 GB	DOJ Executive Office for Immigration Review
ice	17.8M	5	2.0 GB	Deportation Data Project (FOIA litigation)
fec	269.0M	10	29.0 GB	Federal Election Commission
clinicaltrials	56.4M	48	5.8 GB	AACT / ClinicalTrials.gov

Registry format

The registry.json file contains metadata for all databases. Each entry includes:

• id -- short identifier used by the Python package
• name -- human-readable name
• description -- what the database contains
• rows, tables, size_gb -- scale information
• source, source_url -- original data source
• github -- build repository
• huggingface -- Hugging Face dataset page
• attach_url -- direct URL for DuckDB remote attach
• maintainer -- who maintains this database
• license -- data license
• updated -- last update date

Contributing

See CONTRIBUTING.md for how to add a new database to the registry.

Links

• Python package: pypi.org/project/datapond
• Website: datapond-db.github.io/website
• GitHub org: github.com/datapond-db

License

This registry is licensed under the MIT License. Individual databases have their own licenses as specified in the registry.