Statistical Units

Units defined for reporting statistics. Statistical tables are usually published per unit ID without geometry: a population figure for NUTS region DK01, an indicator per LAU. The geometry is borrowed from a reference realisation of the unit, at whatever scale suits the map.

Sphere: Reference Framework

One identity, many geometries

Eurostat publishes NUTS for each version (2016, 2021, 2024) at five scales, 1:1M, 1:3M, 1:10M, 1:20M and 1:60M, all sharing the same NUTS_ID. LAU is published yearly at 1:1M. IDs can change between versions, so a statistical table must state which version its IDs refer to.

Scope

• NUTS (Nomenclature of Territorial Units for Statistics): levels 0–3 for the EU and EFTA
• LAU (Local Administrative Units): the municipal level below NUTS 3; in Denmark the municipalities
• Census tracts and enumeration areas: national units such as the UK’s census output areas
• Statistical grids: square cells used for reporting, such as Danmarks Statistik’s km² grid and the INSPIRE/Eurostat 1 km grid

Statistical grids: predefined and ad hoc

A grid of equal cells makes zones that do not change when administrative borders do, and that can be compared between places and years. Two grids are predefined for Denmark:

• Det Danske Kvadratnet (the Danish square grid), defined in 2001 by Kort- og Matrikelstyrelsen (now Klimadatastyrelsen) and Danmarks Statistik. Square cells from 100 m to 100 km in EUREF89 UTM zone 32 (EPSG:25832). A cell is named by its size and its lower-left corner, truncated to the cell size: the 1 km cell containing N 6188431, E 709802 is 1km_6188_709. A coordinate therefore gives its cell without an overlay, and the cells nest: one 10 km cell holds a hundred 1 km cells. Danmarks Statistik publishes grid statistics on it. Source: Dataforsyningen.
• The European statistical grid (GISCO/INSPIRE), from Eurostat: square cells of 1, 2, 5, 10, 20, 50 and 100 km in ETRS89-LAEA (EPSG:3035), covering Europe, with census population for 2006, 2011, 2018 and 2021. Source: Eurostat GISCO grids.

The two grids do not line up: they are square in different projections, so a Danish cell and a European cell of the same size cut across each other. Choose the one whose statistics you compare with.

An ad hoc grid is generated for the study (generate grid): a cell size or shape no predefined grid offers, such as 250 m squares or hexagons. Recommendations:

• If a predefined grid fits the question, use it, or generate a grid that is aligned with it exactly: the same CRS, an origin on its corner values, and a cell size that divides into its cells. Results can then be compared with official statistics and other studies, and summed to the predefined cells.
• If the cells must be hexagons — for neighbourhood, movement or connectivity — accept that they cannot be compared cell by cell with official statistics; see Square and hexagonal grids.
• Record the cell size, shape, origin and CRS: they are choices that shape the result (the scale and zoning effects of the Modifiable areal unit problem).

Leaves

Classical Theme Mappings