7cd40a6e8e
A pin's position is generalised, so it does not tell a reader the locality or where a group is
centred.
- pins.sharedPoint in all 11 locales and ApproxPinNote.vue's fallback: the groups sharing the point
are named and counted ("{n} groups share this dot"); the locality claim is gone.
- keyCities map.note, map.sheetNote and list.caveat.row in all 11 locales: a cluster is of
people-group pins, not centroids, and each pin marks a people group, not a town.
- The key-cities and languages-per-city caveats in build-tabulator-tables.mjs and in every committed
copy of those two tables (only that one field changes), derive-city-candidates.mjs's caveat, and
the matching comments in key-cities, name-clusters and the poster generator.
- check-approx-pin-note's pass line no longer calls pins localities. Two UI-strings baseline entries
whose strings no longer exist in their files are dropped.
139 lines
6.8 KiB
JavaScript
139 lines
6.8 KiB
JavaScript
// derive-city-candidates.mjs — the LOCATION SPINE for the DX cities corpus.
|
|
//
|
|
// node derive-city-candidates.mjs -> out/city-candidates.json
|
|
//
|
|
// ═══ READ THIS BEFORE TRUSTING THE OUTPUT ═══════════════════════════════════
|
|
//
|
|
// THESE ARE NOT CITIES. They are CLUSTERS OF PEOPLE-GROUP CENTROIDS, and the
|
|
// distinction is the whole integrity of this dataset.
|
|
//
|
|
// The flight brief says "list the key cities where the pens are". Measured against
|
|
// the real pin data, that instruction cannot be followed literally:
|
|
//
|
|
// MEASURED 2026-08-01 (out/_records.json, 2,106 records, all with lat/lng):
|
|
// India 855 pins -> 625 distinct ~11km cells
|
|
// China 154 pins -> 149
|
|
// Sudan 105 pins -> 93
|
|
// Brazil 103 pins -> 102
|
|
//
|
|
// Pins are almost entirely DISTINCT locations, because each one is a people
|
|
// group's centroid — not a settlement. Clustering them into "cities" would
|
|
// manufacture placenames that do not exist, and a fabricated placename in a
|
|
// persecution-sensitive dataset is exactly the class of error that gets someone
|
|
// hurt. So this script does NOT name cities. It emits ANONYMOUS LOCATION
|
|
// CLUSTERS with coordinates, and leaves naming to a cited external source.
|
|
//
|
|
// WHAT IS DEFENSIBLE: at a ~25km radius, 277 cells hold MORE THAN ONE people
|
|
// group. A place where several distinct groups live within 25km IS a real
|
|
// population centre — that inference is sound, and it is the only inference this
|
|
// file makes. Single-pin cells are emitted too, flagged `shared:false`, so a
|
|
// consumer can weight them differently.
|
|
//
|
|
// EVERY row therefore carries provenance and an explicit unknown for anything not
|
|
// derived from the source data. Nothing here is researched; nothing is guessed.
|
|
// Constraint: unknowns are marked unknown; a yes/no is never guessed.
|
|
//
|
|
// PERSECUTION-SAFETY: this file emits location + language + population only. It
|
|
// carries NO gospel-resource, church, or believer field, and must not gain one —
|
|
// that join would be a targeting list.
|
|
|
|
import { readFileSync, writeFileSync, mkdirSync, existsSync } from 'node:fs';
|
|
import { fileURLToPath } from 'node:url';
|
|
import { dirname, join, resolve } from 'node:path';
|
|
|
|
const __dirname = dirname(fileURLToPath(import.meta.url));
|
|
const OUT = join(__dirname, 'out');
|
|
const RECORDS = resolve(__dirname, '..', 'posters', 'out', '_records.json');
|
|
|
|
// ~25km at the equator, narrowing with latitude. Chosen because it is the radius
|
|
// at which multi-group cells appear at all (277 of 1,646); at ~11km almost every
|
|
// pin is alone, and at ~50km unrelated valleys merge.
|
|
const CELL = 4; // quarter-degree
|
|
const cellKey = (lat, lng) => `${Math.round(lat * CELL) / CELL},${Math.round(lng * CELL) / CELL}`;
|
|
|
|
const raw = JSON.parse(readFileSync(RECORDS, 'utf8'));
|
|
const rows = Array.isArray(raw) ? raw : (raw.records || raw.data || []);
|
|
|
|
const cells = new Map();
|
|
for (const r of rows) {
|
|
const lat = Number(r.latitude), lng = Number(r.longitude);
|
|
if (!Number.isFinite(lat) || !Number.isFinite(lng)) continue;
|
|
const k = cellKey(lat, lng);
|
|
if (!cells.has(k)) cells.set(k, []);
|
|
cells.get(k).push({ ...r, _lat: lat, _lng: lng });
|
|
}
|
|
|
|
const clusters = [];
|
|
for (const [key, members] of cells) {
|
|
const lat = members.reduce((a, m) => a + m._lat, 0) / members.length;
|
|
const lng = members.reduce((a, m) => a + m._lng, 0) / members.length;
|
|
const pop = members.reduce((a, m) => a + (parseInt(m.population, 10) || 0), 0);
|
|
const langs = [...new Set(members.map((m) => m.primary_language_label).filter(Boolean))];
|
|
const iso3 = members[0].country_code;
|
|
|
|
clusters.push({
|
|
cluster_id: `loc-${iso3.toLowerCase()}-${key.replace(/[.,-]/g, '_')}`,
|
|
iso3,
|
|
country: members[0].country_code_label || iso3,
|
|
region: members[0].wagf_region_label || 'Other',
|
|
lat: +lat.toFixed(4),
|
|
lng: +lng.toFixed(4),
|
|
// TRUE when >1 distinct people group sits within ~25km — the only signal here
|
|
// that actually argues "this is a population centre".
|
|
shared: members.length > 1,
|
|
people_group_count: members.length,
|
|
people_groups: members.map((m) => m.name),
|
|
languages: langs,
|
|
// Summed from the source records. NOT a city population — it is the combined
|
|
// people-group population within the cell, which is a different quantity.
|
|
upg_population_sum: pop,
|
|
|
|
// ── RESEARCH FIELDS — all unknown until individually sourced ─────────────
|
|
place_name: null, // must come from a cited gazetteer, never inferred
|
|
place_name_source: null,
|
|
internet_access: 'unknown', // yes | partial | no | unknown
|
|
internet_note: null,
|
|
mobile_coverage: 'unknown',
|
|
censorship: 'unknown',
|
|
sources: [],
|
|
|
|
provenance: {
|
|
derived_from: 'doxa people-group records (lat/lng centroids)',
|
|
method: `~25km cell aggregation (quarter-degree), cell=${key}`,
|
|
caveat: 'NOT a settlement. A cluster of people-group pins; place_name is unresolved.',
|
|
derived_on: '2026-08-01',
|
|
},
|
|
});
|
|
}
|
|
|
|
clusters.sort((a, b) => b.people_group_count - a.people_group_count);
|
|
|
|
const byCountry = {};
|
|
for (const c of clusters) byCountry[c.iso3] = (byCountry[c.iso3] || 0) + 1;
|
|
|
|
const stats = {
|
|
source_records: rows.length,
|
|
location_clusters: clusters.length,
|
|
multi_group_clusters: clusters.filter((c) => c.shared).length,
|
|
single_group_clusters: clusters.filter((c) => !c.shared).length,
|
|
countries: Object.keys(byCountry).length,
|
|
cell_size_deg: 1 / CELL,
|
|
named_places: clusters.filter((c) => c.place_name).length,
|
|
internet_researched: clusters.filter((c) => c.internet_access !== 'unknown').length,
|
|
generated: '2026-08-01',
|
|
scope_note: 'Scope is the full 2,106-record DX people-group corpus. No separate "2000 list" artefact exists on disk — verified by search — so this corpus IS the scope.',
|
|
safety_note: 'Contains NO gospel-resource, church or believer field, and must not gain one. That join is held pending its own architecture (separate store, never joined to UPG locations in a public artifact).',
|
|
};
|
|
|
|
if (!existsSync(OUT)) mkdirSync(OUT, { recursive: true });
|
|
writeFileSync(join(OUT, 'city-candidates.json'), JSON.stringify({ stats, clusters }, null, 1));
|
|
|
|
console.log(`[cities] ${stats.location_clusters} location clusters from ${stats.source_records} records`);
|
|
console.log(`[cities] multi-group (real population centres): ${stats.multi_group_clusters}`);
|
|
console.log(`[cities] single-group: ${stats.single_group_clusters}`);
|
|
console.log(`[cities] countries: ${stats.countries}`);
|
|
console.log(`[cities] named: ${stats.named_places} internet-researched: ${stats.internet_researched} (both 0 until sourced)`);
|
|
console.log('[cities] top multi-group clusters:');
|
|
for (const c of clusters.slice(0, 8)) {
|
|
console.log(` ${String(c.people_group_count).padStart(2)} groups ${c.country.padEnd(28)} ${c.lat},${c.lng}`);
|
|
}
|