Methodology

1. Approach & assumptions

Every unit in this dataset sits deliberately between two administrative scales that already exist: NUTS 0 (the nation-state) and the official NUTS 1 macro-region. We call the units in between practical regions — not a new layer of government, but a candidate map of where cultural-economic coherence already exists, whether or not any current boundary draws it that way.

Why non-national units

This dataset is not a re-slicing of the nation-state at a smaller grain. It pursues two aims together: cultural unity — the cultural-units layer described below, identity-true at whatever grain that requires — and functional self-governance, the devolution and economic-self-sufficiency case the autarky index (further down this section) exists to measure. Germany is the model worth naming directly: it already operates as at least three centers of gravity with their own economic potential — Berlin, Munich, and Cologne/the Ruhrgebiet — rather than one dominant capital radiating outward. The asymmetry this dataset argues against is the current one, where a small state like Slovenia sits formally equal but practically outweighed next to a state many times its population and voice; the fix is not to make Slovenia bigger, but to work at the scale of far smaller regions everywhere. That rescaling only makes sense together with a strong European identity in its own right — without one, nothing actually changes: national leaders simply become the layer everything gets renegotiated through again, bypassing whatever the regional map says.

Input layers

Each region was assembled by triangulating several independent, often disagreeing layers of evidence, rather than any single official source:

• Language and dialect — recognized regional languages and dialect-atlas classifications (e.g. Pellegrini's Carta dei dialetti d'Italia, Wiesinger's East Franconian classification, Kashubian/Silesian/Moravian census self-declarations).
• Administrative reference — current NUTS-1/2/3 geometry and historic administrative units (duchies, Kingdoms, pre-1990s planning regions) where they carry living identity.
• Watersheds and valleys — river-basin and Alpine/fjord topography that structures settlement (e.g. the Rhône-Alpes corridor, Norway's five fjord-culture nodes).
• Aquatic bioregions (planned layer) — a finer, hydrology-first classification than the topographic watershed/valley signal above: HydroSHEDS-based drainage-basin bioregions, cross-referenced from Jakob's Unlocking Housing project (idea2.site/unlocking-housing). Not yet integrated into any region-boundary decision in this dataset; planned as a dedicated lens in the map's perspectives panel (alongside the political, linguistic, and self-sufficiency perspectives) once that geometry is exported for this dataset's scope.
• Commuting basins — functional urban areas, used to decide how far a metropolitan region's gravity extends (e.g. which Home Counties belong to London's commuter belt versus East Anglia).
• Urban cultural centers — the presence of an institution-bearing city (media, universities, regional government) that a region can plausibly organize around.

Inclusion test

A candidate region had to clear a three-part test:

1. one urban center already functioning as a cultural hub;
2. its own diversified industrial base (not a monoculture);
3. a distinct sub-national identity that is dependent on, not seeking independence from, its national state.

Where the input layers genuinely conflicted — a dialect continuum crossing a national border, a minority-language pocket inside a larger region, a contested internal boundary — the conflict was recorded as a toggleable conflict zone (conflict_flags in the data) rather than force-resolved into one boundary. 25 such flags exist across the dataset, from Eupen-Malmedy to Sápmi to Kosovo's internal north.

NUTS-vintage policy

Region nuts_codes follow the GISCO 2024 geometry vintage (NUTS_RG_20M_2024_4326.geojson), not the 2027 xlsx recoding that data/nuts_index.json is drawn from, wherever the two disagree. Three countries' codes were affected: Norway's 2024 county geometry (NO02/06/07/08/09/0A1/0A2/0A3) rather than the 2027 sheet's NO0C/0D/0E; Belgium's 2024-vintage BE21/23 rather than the 2027 recoding's BE26/27; and Bulgaria's pre-reform BG3/BG4 rather than the 2027 sheet's BG5/BG51-54. The rule throughout: author against whichever file the pipeline can actually render, and flag the mismatch rather than silently reconcile it.

Geometry sources

• GISCO NUTS_RG_20M_2024_4326.geojson — EU27, EFTA, and enlargement/candidate countries (default source).
• GISCO NUTS_RG_20M_2021_4326.geojson — United Kingdom (post-Brexit UK boundaries no longer ship in the 2024 release) and, for the single mixed-source ireland-unified record, Ireland too, since the 2021 file happens to carry both countries' geometry under one vintage.
• GISCO CNTR_RG_20M_2024_4326.geojson — country-level polygons for microstates (Andorra, Monaco, San Marino, Vatican City) and for Moldova and Georgia, which have no sub-national NUTS geometry available in this pipeline.
• geoBoundaries-UKR-ADM1 (© OpenStreetMap contributors via geoBoundaries, ODbL) — Ukraine's 27 oblast-level units, fetched because GISCO's own NUTS file ships Ukraine as a single undivided country polygon with no SR2024 sub-national breakdown.

Autarky index

Each region carries a 7-dimension self-sufficiency estimate — Water, Food, Shelter, Energy, Transportation, Hygiene & Healthcare, and Clothing — scored 1–5 with a composite equal to the arithmetic mean. Security and Digital/Connectivity are explicitly out of scope: the index measures material provisioning capacity, not defensibility or network resilience, and folding either in would have required a different (and much more contested) evidentiary basis.

• Water
• Food
• Shelter
• Energy
• Transportation
• Hygiene & Healthcare
• Clothing
• Security (out of scope)
• Digital/Connectivity (out of scope)

Scores are expert-style estimates, not measurements. We are explicit about that subjectivity rather than dressing the numbers up as harder than they are: each score carries a one-to-three sentence rationale citing the decisive factor (e.g. a region's water score reflects watershed/reservoir capacity relative to population; energy reflects generation mix and grid dependency), authored per-region rather than derived from a single external statistical series. Composite scores should be read as relative-ranking signals across the dataset, not as validated indices.

2. Priority axes & re-drawn boundaries (v0.1.2)

The input-layer list in section 1 is not a flat checklist — the user set an explicit priority ordering for this draft, and v0.1.2's boundary changes (splits, merges, re-homings) were made by re-weighing every contested unit against these axes in this order:

1. Culture & عصبية (ʿaṣabiyya) — Gemeinschaftsgefühl / group identity. Ibn Khaldūn's term for the social solidarity that binds a group together comes first: does a population actually feel itself to be one thing, independent of where a modern administrative line happens to run? This is the axis that drove the Styria≠Carinthia, Basel/Bern-Mittelland/ Romandie Grossregion groupings, Upper Silesia's Ślązacy self-declaration, and the Kashubia/Kuyavia/Warmia-Masuria "fit" gradient (section 6).
2. Einzugsgebiete (catchment areas) — the functional area a city or region genuinely draws people, commuters, and services from, independent of identity. Used to decide edge cases once identity alone under-determines a boundary (e.g. Deventer's Overijssel/Achterhoek catchment, Preston-Lancaster's dipole).
3. Train lines — rail corridors as a proxy for lived, everyday connectedness (commuter and intercity), read where identity and catchment both leave a boundary open.
4. Waterways — rivers and canals as historic trade and settlement corridors (the same logic already underlying the Rhône-Alpes and Mur/Enns-valley cases from v0.1.0).
5. Watersheds — drainage-basin topography, the lowest-priority tiebreaker, used mainly for Alpine and fjord-culture cases where no stronger signal resolves the line.

Every v0.1.2 boundary change was re-drawn through this lens stack, highest-priority axis first: ʿaṣabiyya/identity as the primary test, catchment areas and transit/water corridors as tiebreakers where identity alone is ambiguous or contested. Where a v0.1.0 boundary already matched this ordering, it was left untouched — the axes are a re-verification method, not a mandate to re-draw everything from scratch.

3. Two-level model: cultural units & representation groupings

Philosophy: define the cultural atoms first, at the finest grain that still passes the inclusion test (section 1) and the priority-axes re-check (section 2) — then, and only then, group those atoms into a separate representation layer sized for practical governance. The two questions — "where does identity actually run?" and "how many seats can a chamber usefully have?" — are answered independently, in that order, so that a representation constraint never quietly overrides an identity finding.

Loading live counts from the current data build…

Cultural units (fine-grained, identity-true)

The data/regions/*.json zone files are the atomic layer: every record that clears the inclusion test on its own ʿaṣabiyya grounds gets its own unit, regardless of how small that makes it relative to a governance-sized region. This is why the dataset now runs to roughly two hundred units rather than converging early on something closer to the 40–70 target — that convergence happens one layer up, not here.

Representation groupings (40–70 target)

data/groupings.json partitions every default-visible cultural unit into exactly one grouping — singletons are allowed where a unit is already large enough (or too distinct, e.g. a sovereign microstate) to stand alone. This layer, not the cultural-unit layer, is what the original 40–70 target was always meant to size. Content agents working zone-by-zone initially overshot this target substantially (a zone-by-zone artifact, documented candidly in the v0.1.2 changelog); a dedicated cross-zone consolidation pass brings the dataset-wide total toward the target range. Read the number in the stat strip above as authoritative for whichever build served this page, not the prose here.

The map switch

The map exposes a Cultural units ⇄ Representation groups master switch. In cultural-unit mode every atom renders with its own boundary and label. In representation mode, grouping members visually dissolve into their parent grouping — a thick shared outer border replaces the internal seams, and a single label covers the whole grouping — without deleting or renaming the underlying cultural units, which remain the dataset's source of truth.

EP delegation model: population balance across representation groupings

v0.1.2 keeps the European Parliament chamber as-is rather than proposing a new institution — the change is in how seats would be allocated: each representation grouping sends a roughly equal delegation, instead of national delegations sized by degressive proportionality. The design target for the resulting population spread is a max factor of 3–7 between the largest and smallest grouping, 10 at the absolute outside — tighter than today's EP, where Germany's ~96 seats against Malta's 6 already implies a 16× spread by seat count, and the per-seat population gap runs to roughly 10× (Malta ≈86,000 people per seat vs. Germany ≈860,000).

Amendment A3 (round-4 feedback) adds a second grouping layer purpose-built to hit that target, alongside the original affinity-based layer, and Amendment A4.1 (round-5 feedback) adds a third, state-boundary variant: data/grouping_schemes.json carries the original Affinity (66) scheme (culturally-driven, unchanged from the count-target discussion above) plus however many further schemes — such as Balanced, engineered so its non-exempt max factor lands at or below 7×, and A4.1's balanced-national variant, which additionally forbids any grouping from crossing a state border — the current data build defines. Groupings a scheme marks factor_exempt (sovereign microstates, exclaves, and similar singletons) sit outside that scheme's factor calculation and are listed separately as special status, rather than dominating the spread the way they did under the single-scheme analysis. A scheme may additionally mark a sovereign_floor list (A4.1 — small sovereign states that cannot merge upward without crossing a state border): those groupings are excluded from the factor too, but labeled distinctly as "sovereign minimum — cannot merge without crossing a border" rather than "special status", since they represent a structural floor rather than an exemption. Every scheme's live stats are shown side by side below, generated from the data file rather than a fixed pair; the full data-driven breakdown lives on the data table page.

4. Autarky index: per-dimension rationale & beta dimensions

Section 1 already describes the 7-dimension autarky index and its composite score. v0.1.2 makes two additions on top of that unchanged foundation.

Required per-dimension rationale

Every region's autarky.rationales object now carries a required, non-empty, one-sentence score basis for the three dimensions the user singled out as most load-bearing — water, food, and energy — citing the concrete factor behind the number (a watershed, a specialty export, a hydropower/grid-dependency fact). Rationales for the remaining four dimensions are optional and present wherever a region's authoring agent had a decisive factor worth recording. This is additive to, not a replacement for, the single-sentence autarky.rationale field already described in section 1.

Beta dimensions

Four of the seven dimensions are marked beta in this build: shelter, transportation, hygiene & healthcare, and clothing. "Beta" here means the dimension is judged easily importable within the continent — a shortfall is a logistics/trade problem, not a structural one, unlike water or food where geography sets a hard local ceiling. Selecting a non-composite beta dimension in the detail panel shows a small box under the region name with that dimension's score basis and a "beta" badge, so the distinction is visible at the point of use, not only in this document.

Hygiene & Healthcare, redefined

This dimension is redefined for v0.1.2 as the region's industrial capacity to produce the medical and sanitation wares it needs (masks, basic pharmaceuticals, dressings, water-treatment chemicals, hospital consumables) — not bed counts, staffing ratios, or clinical outcomes, which belong to a health-system-quality question this index does not attempt to answer. A region can score low here while having excellent hospitals, if it manufactures none of what those hospitals consume.

Honest subjectivity, unchanged

The section-1 caveat still applies without softening: these remain expert-style estimates, not measurements, read as relative-ranking signals rather than validated indices. Adding required rationales makes the basis for each score legible and checkable — it does not make the scores themselves more objective.

5. Parliament rule & exceptions

Rule: any region with its own parliament gets its own cultural unit. A parliament — a directly elected legislative body with genuine law-making competence — is treated as the single strongest available proof that a population already governs itself as one thing, stronger than dialect maps or catchment analysis alone. Applying this rule swept every in-scope state: it resolved several v0.1.0 parsimony merges that had been flagged as open tensions (Rhineland-Palatinate≠ Saarland, Berlin≠Brandenburg, Styria≠Carinthia, Salzburg≠Upper Austria), and it is why Bosnia and Herzegovina's three entity parliaments (Republika Srpska, the Federation, Brčko) each now get their own unit.

The rule is not applied blindly, though: most of Europe's unitary states (France, Netherlands, Nordics, Ireland, Poland's, Czechia's and Slovakia's sub-national tiers, most of Southeastern Europe) have no regional parliament at all below the national legislature — county/kraj/voivodeship/oblast councils are deconcentrated administration, not legislatures, and the rule simply does not fire there. The exceptions below are the opposite case: units where a real, legislative sub-national parliament does exist, but the cultural unit was deliberately kept spanning more than one of them anyway — the user asked explicitly to see these, rather than have them absorbed silently into the "it's fine, don't worry about it" pile.

Rendered client-side from data/parliament_exceptions.json — the same file the pipeline and app both read, so this table cannot drift from the dataset it describes.

The inverse case: one parliament, several cultural units

The exceptions table above covers units that merge several real parliaments. The opposite pattern also exists in the dataset, and is just as real an exception to a naive "one parliament, one unit" reading, even though it doesn't fit the merge-table's columns: a single real parliament whose jurisdiction is split across several cultural units, because the identity case for splitting outweighed the administrative case for one parliament, one region. This affects Baden-Württemberg (one Stuttgart Landtag, split into de-baden and de-wuerttemberg on their distinct pre-1952 histories), Bavaria (one Bayerischer Landtag in Munich, split into de-franconia, de-bavarian-swabia and de-munich-altbayern), North Rhine-Westphalia (one Düsseldorf Landtag, split into de-rhineland and de-westphalia), and Scotland (one Holyrood, split into uk-scotland-central-belt and uk-scotland-highlands). A related but distinct partial case: the Sámi Parliaments of Finland, Sweden and Norway are real elected bodies with genuine cultural-autonomy competence, but not territorial legislatures over their whole host region (fi-lapland-sapmi, se-norrland-upper, no-tromso-north-norway) — a partial exception worth naming even though it fits neither table direction cleanly.

6. Variant sets: alternative views

Some regions are genuinely contested between two or more defensible boundary readings — not a case where one answer is right and the rest are mistakes, but one where the honest move is to let the user see and switch between the readings directly. v0.1.2 introduces … variant sets (data/variant_sets.json) for exactly this.

The three-view pattern

Most sets follow the same three-view shape, though the exact labels are set-specific:

• Current — the region(s) as they sit within today's state boundaries (the dataset's default rendering).
• Connected (sometimes "unified") — a hypothetical cross-border or cross-unit merge that follows the cultural/historical case to its conclusion, ignoring the state line.
• Overlap — both readings shown at once, translucent, with a hatch pattern where they physically overlap, so the disagreement itself becomes visible rather than adjudicated away.

On the map, each set renders a small clickable chip at its authored anchor coordinate (zoom-aware); clicking cycles or opens the view options directly at the point on the map the set concerns, rather than burying the choice in a sidebar list (the sidebar keeps a compact mirror for accessibility/search, but the map chip is the primary interaction — this was explicit user demand).

Notable calls

• Basque over-coverage. The connected view's French-side geometry (FRI15, Pyrénées-Atlantiques) covers all of Béarn, a distinct non-Basque territory, not just Iparralde — an honest over-coverage, stated directly in the view's own note rather than smoothed over. The fix (geoBoundaries FRA ADM3, Bayonne-arrondissement geometry) is identified but not yet integrated; planned for a future pipeline pass.
• Northern Ireland stays distinct in every view of the ireland set, including the all-island unified view — it renders as a distinct outlined unit on top of the united-Ireland shape rather than disappearing into it, per explicit user instruction that a unification frame must not erase the live internal fault line it would actually have to resolve.
• Dutch lowlands, explicitly hypothetical. The flanders set's connected view (dutch-lowlands-unified) merges Flanders with the Dutch language area across the Belgian-Dutch border; it is labelled as a cultural/linguistic hypothetical throughout, not a governance proposal — component units keep their real parliaments (Vlaams Parlement, Staten-Generaal) in every other view.
• Silesia, Poland+Czechia. The silesia set's current view keeps Poland's Upper Silesia (Katowice) and Czech Silesia separate by state; its connected view reunites the historic Duchy of Silesia across the modern border as silesia-unified — a cultural mapping exercise, explicitly not a territorial claim.
• Liverpool/Manchester overlap. Per direct user request ("suggest a variant with overlaps especially for Liverpool and Manchester"), the two city-regions get a dedicated overlap view showing both catchments translucent, without inventing an artificial third "shared belt" region.
• Pomerania, two versions. The default broad view matches Statistics Poland's own PL6 macroregion; the split view exposes the honest fit gradient behind that convenience, per the user-supplied table:

7. The six push-back challenges

The project brief itself mandated that six specific claims be challenged rather than assumed. Each is addressed here with the verdict reached and the evidence behind it.

8. Explicit disagreements with the spec's fixed decisions

The brief instructed that fixed decisions be applied and verified, but challenged rather than silently overridden where the evidence pointed elsewhere. These are the live disagreements on record:

9. NUTS-1 comparison

Loading… of the dataset's default-visible cultural units explicitly diverge from the official NUTS-1 (or, for non-EU/non-Eurostat countries, the corresponding SR-1) boundary they overlap — figures reflect whichever build served this page, not a number frozen at v0.1.0's 109-of-191. The table below is rendered client-side from data/regions.json and data/nuts1_reference.json: for every region with a non-empty nuts1_disagreement, we derive its overlapping official NUTS-1 unit(s) by taking the first three characters of each of its nuts_codes (the NUTS-1 code is always the country-plus-one-character prefix of any finer NUTS code by construction) and looking each one up against the reference list, mapping Greece's regional code (GR in this dataset) to its official Eurostat code (EL). Where no reference entry exists — the United Kingdom (excluded from Eurostat's NUTS classification entirely since Brexit) and the CNTR-sourced microstates/Moldova/Georgia (which have no NUTS-1 subdivision to compare against) — the cell says so rather than guessing.

Reverse summary: how official NUTS-1 units fare

Computing…

10. Data limitations

11. Sources

Sources are tiered: A legal/government/ statistical-office primary sources, B journalism and analysis, C community and dialectology-adjacent sources. Sources older than ten years carry a ⚠️ flag.

• A Eurostat NUTS / GISCO reference geometry and classification (NUTS_RG_20M_2024, NUTS_RG_20M_2021, CNTR_RG_20M_2024)
• A EU Functional Areas toolkit (commuting-basin definitions)
• A Destatis (Federal Statistical Office of Germany)
• A INSEE (France), Statbel (Belgium), CBS (Netherlands)
• A ISTAT (Italy), INE (Spain), Statistics Portugal / INE Portugal
• A ONS, NISRA, CSO Ireland (UK/Ireland regional statistics)
• A GUS (Poland), Czech Statistical Office, ŠÚ SR (Slovakia)
• A KSH (Hungary), including regscience.hu regional-policy analysis
• A State Statistics Service of Ukraine (UkrStat), pre-war/pre-occupation baselines
• A 2021/2011 national censuses (Poland — Silesian nationality; Czechia — Moravian nationality) ⚠️ 2011 entry >10 years
• A geoBoundaries (Ukraine oblast ADM1 geometry) — © OpenStreetMap contributors, ODbL
• B Assembly of European Regions (2021) — Hungarian regionalisation policy analysis
• B Regional press and regionalism analyses (Scouse/Merseyside devolution, Cornish national-minority recognition, Norrlandsfrågan, Szlonzoki self-declaration reporting)
• C Pellegrini, G.B. (1977) — Carta dei dialetti d'Italia (Emilia/Romagna dialect classification) ⚠️ >10 years
• C Wiesinger, P. (1983) — East Franconian dialect classification ⚠️ >10 years
• C Council of Europe Framework Convention for the Protection of National Minorities (2014 Cornish recognition) ⚠️ >10 years
• C Community/dialectology sources for regional-language status (Kashubian, Rusyn, Occitan-Catalan Roussillon)
• B Ibn Khaldūn, Muqaddimah (14th c.) — origin of ʿaṣabiyya as a social-solidarity concept, used here as a priority-axis label rather than a historical citation ⚠️ >10 years
• A Sametinget (Sweden), Sámediggi (Norway), Saamelaiskäräjät (Finland) — Sámi Parliament competence and jurisdiction, for the parliament-rule partial-exception cases
• A Statistics Poland (GUS) "Polska Wschodnia" macroregion definition; Swiss Federal Statistical Office (BFS) Grossregionen definitions — official-macroregion precedents cited in the parliament-exceptions table
• A GISCO CNTR_RG_20M_2024 — Armenia (AM), Guernsey (GG), Jersey (JE), Isle of Man (IM), Gibraltar (GI) country-level geometry, all already present in the committed file (no new fetch required)

Pellegrini (1977) and Wiesinger (1983) are both older than ten years ⚠️ and are used here for historical dialect classification, where the underlying linguistic boundaries change on a much longer timescale than the sourcing recency convention assumes — flagged rather than silently exempted.

12. Kaliningrad note

13. Changelog — v0.2.2 draft

v0.1.3 (this build, Amendment A4): contested/cultural overlays (Sápmi, Székely Land, Kashubia, and similar) are now off by default behind a "Contested & cultural overlays" master toggle in the map sidebar (A4.2), and data/grouping_schemes.json gains a state-boundary-respecting balanced-national scheme with a sovereign_floor list for states that cannot merge upward without crossing a border (A4.1). The itemized list below documents the earlier v0.1.0 → v0.1.2 diff.

A concise list of what changed. Where a figure is cheap to compute from the live data build, it is pulled client-side and marked (live); everything else is a fixed count as of the v0.1.3 draft build, not re-verified on every page load.

• Two-level model. Cultural units (identity-fine) and representation groupings (40–70 target) are now separate layers — see section 3 — resolving the count tension section 7(f) originally documented as unresolved.
• Coverage holes closed. Spain (La Rioja, Cantabria, Asturias, Extremadura, Murcia, Ceuta+Melilla), Portugal (Alentejo, Centro), Italy (Basilicata/Potenza), and Greece's Ionian islands are all now covered — every one of these was explicitly listed as a v0.1.0 coverage gap (section 7's push-back challenge (e), "assessed and excluded" paragraph, kept above for the historical record — now superseded, see that section's own v0.1.2 note) and is closed as of this build.
• New landmass/territory additions. Armenia (am-armenia), the Channel Islands (Guernsey and Jersey, as two separate units — each Bailiwick has its own legislature), the Isle of Man, Gibraltar, and East Thrace (tr-east-thrace, the mandated-covered landmass west of Istanbul, with the rest of Turkey explicitly out of scope).
• Ukraine historic splits. The v0.1.0 central-Ukraine catch-all is dissolved into three historically-grounded units: the Cossack Hetmanate core (Poltava+Cherkasy+Chernihiv), Podillia (Vinnytsia+Khmelnytskyi), and Volhynia (Volyn+Rivne+Zhytomyr); Kirovohrad is re-homed to the Dnipro-Zaporizhzhia unit on Zaporozhian settlement-history grounds, reasoned in that region's record.
• Slovakia dipole. Superseding an earlier finer split, Slovakia is now exactly two units: a Bratislava-and-western- Slovakia capital region, and a Košice-Prešov eastern dipole on a Rusyn/Ukrainian-minority basis.
• Rijeka/Kvarner becomes its own Croatian unit (Zagreb/Continental, Istria, Rijeka, Dalmatia — four units where v0.1.0 had fewer).
• Székely Land carved out of Transylvania as its own visible cultural unit (Covasna+Harghita+Mureș), while remaining outside the parliament rule's exception table since no legally-recognized Székely assembly currently exists (section 5 explains the distinction).
• Silesia and Pomerania variant sets — see section 6 — expose the Poland+Czechia cross-border Silesia reading and the Kashubia/Kuyavia/Warmia-Masuria fit gradient behind the default "broad Pomerania" convenience.
• Poles and reach. Every unit now carries a pole classification (monopole/dipole/polycentric) with a central_coord map marker, and dipoles carry a secondary/secondary_coord pair too (rendered as filled/hollow dots respectively); some units additionally carry a reach array documenting cross-border or cross-region functional influence (Munich, Graz into Slovenian/Carinthian areas, capitals generally).
• Beta autarky dimensions — shelter, transportation, hygiene & healthcare, and clothing — are now explicitly labelled as easily importable within the continent, with hygiene & healthcare redefined as production capacity for medical/sanitation wares rather than clinical-quality proxy (section 4).
• Sweden's Norrland split at the Sápmi line: Upper Norrland (Umeå/Luleå) is distinguished from Lower Norrland (Sundsvall/Östersund), mirroring the same-line splits already modeled for Finland and Norway.
• Special-status (dashed) styling is now a first-class schema field, applied to enclaves, microstates, capital city-states, and de-facto-divided or historically-anomalous units (Brussels, Vatican City, Liechtenstein, Vorarlberg's dual orientation, Andorra, Gibraltar, San Marino, Hamburg, Monaco, Ceuta+Melilla, Piedmont's Occitan/ Franco-Provençal valleys, the Crown Dependencies, Cyprus's de-facto division, and East Thrace) — see the live count above.