Skip to content

[FEATURE] Use the published Insgesamt instead of summing perturbed census categories #1530

Description

@jh-RLI

Zensus 2022 protects published figures with the Cell-Key method, which perturbs each published figure independently. The method paper advises explicitly against "eigenständige Summenbildung aus überlagerten Fallzahlen statt auf das bereitgestellte überlagerte Insgesamt direkt zuzugreifen" — deriving a total by summing perturbed categories instead of reading the published total.

The conversion prepared for #1374 does exactly that for HEIZTYP: it sums the six categories and discards the supplied Insgesamt_Heizungsart. Measured over 2,628,942 cells, the two differ in 45.4 % of cells and by 1.11 % in aggregate.

This is not a defect. For the ratio computed in district_heating_areas, numerator and denominator then come from one consistent set, which is arguably better than mixing a category with a separately perturbed marginal. But it is a conscious deviation from the data provider's guidance, and it should be revisited deliberately rather than by omission.

It touches the pre-processing chain, so it pairs naturally with the heating energy carrier work.

Refs #1240, #1374.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

No labels
No labels

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions