If you need official statistics from Rwanda or Kenya in code, start here. We publish the same versioned data through a Python package, a REST API, and the catalog, so you can use the route that fits the work in front of you.
Choose your path
You do not need to learn the whole system before loading a file. Pick the route that matches what you are trying to do, then use the reference when you need a detail.
| If you want to | Start here | Why this route |
|---|---|---|
| Load a dataset into a DataFrame now | Your first dataset | Installs the package and loads a real series in 5 steps |
| Use pandas or Polars in an existing project | Load data | Shows loading, dates, caching, metadata, and offline behavior |
| Work in SQL without the package | Query with DuckDB | DuckDB reads our Parquet files directly over HTTPS |
Check an argument to imi.load() | Python package reference | Lists every function, parameter, result, and error |
| Call the service from another language | REST API reference | Documents metadata, downloads, previews, search, and status |
Understand an ID such as rw.nisr.cpi.monthly | Dataset IDs | Explains every segment and the permanence rules |
How the pieces fit
We maintain 1 catalog of dataset records. Each record tells you what the data is, who published it, where the source lives, what its publisher says about reuse, and which version you received.
- Each dataset record holds the source, schema, license, and version history.
- From the record we publish data files, which both the Python package (
imibare) and the REST API (api.imibare.org) serve. - The catalog at imibare.org is built from the same records and reads its previews through the API.
The catalog is the best place to see current coverage. The package and API are the better places to take the data into your own work.
Use data with its context
We did not create the underlying statistics, and we do not grant a blanket license for them. Each dataset page records what its publisher says, including when the publisher says nothing, and links to the original publication.
When you publish work that uses a dataset, cite both the institution that produced it and the specific imibare version you used. Read the terms of use when you need the complete position on rights, attribution, and limitations.