Skip to content

imibare documentation

Government statistics, programmatically. Free, versioned, open.

imibare makes official government statistical data programmatically accessible as a Python package, a REST API, and a browsable catalog site. Every dataset ID is country-prefixed, so more countries slot in without breaking changes; the catalog lists what is covered today.

System overview

graph TD
    Y["pipeline/catalog/datasets.yaml<br/>Single source of truth"] --> P["Dagster ETL pipeline<br/>pipeline/assets/"]
    P --> R[("Cloudflare R2<br/>imibare-data bucket")]
    R --> PKG["Python package<br/>pip install imibare"]
    R --> API["REST API<br/>api.imibare.org"]
    Y --> SITE["Catalog site<br/>imibare.org"]
    API --> SITE

All three products are driven by a single file: pipeline/catalog/datasets.yaml.

Who is this for?

I want to…Start here
Load a dataset into a DataFrame right nowQuickstart tutorial
Use the Python packageHow to load data
Query datasets with DuckDBHow to query with DuckDB
Look up imi.load() parametersPython package reference
See all REST API endpointsREST API reference
Understand the dataset ID schemeDataset IDs reference

Documentation structure

These docs cover getting government data into your tools:

Browse the datasets

Every dataset lives in the imibare catalog, with search, filters by institution and topic, and a page per dataset showing its columns, coverage and source. The catalog is the authoritative list; these docs describe how to get the data into your tools.

To list datasets from code, use imi.catalog() or the /datasets endpoint.

Licenses

Data: terms of use

imibare claims no copyright in the data it republishes, and makes no representation about what the publishing institutions may claim. Each dataset carries a license recording what its own publisher states, which for most of the catalog is nothing at all, alongside a license_basis saying how that was established. Cite both the source institution and the imibare version you used.