mloda-plugin-govdata

German open government data as typed tables you can trust: found, downloaded, parsed, made comparable, and traceable to its source. A plugin for mloda: you name the columns, it does the rest.

Why

The data is open, but every team rewrites the same scripts:

• German CSV. Decimal commas, multi-row headers, and signs like - and . that mean different things.
• Changing borders. Göttingen and Osterode merged in 2016. The raw Destatis table shows three districts, and zeros for the years each one did not exist.
• Lost provenance. Which file, which license, which attribution: gone by the time the chart is done.

What's inside

Install

pip install mloda-plugin-govdata

Quick start

No account needed for GovData:

from mloda.user import Feature, mloda
from mloda_plugin_govdata.feature_groups.govdata import StuttgartPopulationReader

slug = "einwohner-nach-altersgruppen-und-stadtbezirken"
result = mloda.run_all(
    [Feature("Einwohner", options={StuttgartPopulationReader: slug})],
    compute_frameworks=["PyArrowTable"],
)

Destatis, with a free GENESIS account:

from mloda_plugin_govdata.feature_groups.destatis import DestatisReader

result = mloda.run_all(
    [Feature("value", options={DestatisReader: "12411-0010"})],  # population by Land
    compute_frameworks=["PyArrowTable"],
)

Göttingen on today's borders, with value__rebased (full example):

~key   ~year  ~value    ~flag     ~sources
03159  2013   322616.0  rebased   [03152, 03156]
03159  2015   329538.0  rebased   [03152, 03156]
03159  2016   327065.0  observed  []

Demo

A marimo notebook walks through every source, the Göttingen series, and a join of population and eligible voters for all 16 Länder:

git clone https://github.com/mloda-ai/mloda-plugin-govdata.git
cd mloda-plugin-govdata
uv sync --all-extras
uv run marimo edit demos/govdata_demo.py

The Destatis chapters skip themselves without GENESIS credentials. A second notebook, demos/berlin_wahl_2026_demo.py, reads four Berlin elections with one set of reader options.

Docs

• Readers: every source, dataset search, peek
• Destatis options and credentials
• Harmonization
• Recipes
• Architecture and adding a reader

Status

Young but working, built during a Prototype Fund stage (June to November 2026). The API may still change; the release notes give a migration for every breaking change.

Funding

Funded by the German Federal Ministry of Research, Technology and Space (BMFTR) through the Prototype Fund, supported by the Open Knowledge Foundation Deutschland. Funding code: 16IS26S11.