How it works

The pipeline

A program fetches the public data, computes every figure in advance and saves it as small files; the site only reads those files. In order: open-data APIs → a Python/pandas pipeline → committed data/*.json → a static JS frontend → Firebase Hosting. The pipeline is the only thing that touches the upstream APIs; each source is an isolated module, so a failure in one can't break the others. A scheduled job re-runs it twice a month and commits the refreshed JSON.

Choices that apply everywhere

Energy data is easy to get subtly wrong, so each view states the choices behind its figures. Each derived figure also has its own method note stating inputs, weighting, window and known limits. These choices apply everywhere:

• Prices are averaged to whole hours. Germany moved to quarter-hour settlement in October 2025, and quarter-hour spreads are wider, so hourly spreads are a lower bound.
• All times are German time (CET/CEST), including the 23- and 25-hour days when the clocks change.
• The battery-arbitrage and flexibility figures are labelled upper bounds: they assume perfect foresight and no losses, so real revenue is lower.
• Carbon intensity is counted where the power is generated, with lifecycle factors. Gaps in the source show as gaps in the chart. Negative prices and residual load are kept at their full values.
• Where a metric needs a threshold or a different methodology, the choice is stated in the view itself.

Method written and checked by Sebastian Knödel.

Working on something that needs this data? Say what you need.