METHODOLOGY · updated June 2026

How AJG Global Nexus composes the unified knowledge graph

The architecture, composition rules, data sources, and editorial standards behind 237,718 cross-referenced entities across trade, policy, cities, and global commerce.

1. Philosophy · composition over syndication

AJG is a deterministic composition engine, not a syndicator. Every page on this site is generated from a curated entity graph using type-specific templates. We never scrape and republish third-party content. We never rewrite news articles. The content you read on AJG is our own composition, anchored in public-domain data points (country codes, continents, tiers, categories, trade corridors, institutional definitions) and cross-referenced through the graph.

This matters for three reasons: it's copyright-safe, it scales freely (adding 1,000 cities doesn't require 1,000 editorial hires), and it's differentiated — there is no competitor on the internet doing exactly this.

2. The unified knowledge graph · 7 entity types

Every thing on AJG is one of seven entity types, each with its own ID schema, URL space, and department views:

TypeCountKey schemaSchema.org @type
City234,311city::{slug}Place
Topic2,890topic::{slug}DefinedTerm
Scope lens83scope::{slug}DefinedTerm
Desk feed68desk::{slug}Dataset
Library category50library::{slug}Collection
Tool15tool::{slug}SoftwareApplication
Lexicon term116lexicon::{slug}DefinedTerm
Total237,718× 10 clean department URLs each = 2,377,180 URL surfaces

3. Sources · primary, auditable, public-domain

Our data layer is built from public-domain reference material: country and city lists from ISO 3166-1/2 and UN M.49; continent groupings from UNSD; city tier classifications from urban research literature; population figures from UN World Urbanization Prospects; trade taxonomy from the WCO Harmonized System; Incoterms from ICC 2020; international body memberships from official secretariat websites; lexical definitions synthesised from primary regulatory texts (EU directives, US statutes, trade agreements).

No private APIs. No scraped third-party content. No runtime dependencies. Every data point is traceable to a public-domain source.

4. Composition rules · how pages are generated

Each entity hub (e.g. /cities/mumbai/) is assembled at render time from:

  1. The entity's base data (name, description, country, continent, tier, category, tokens)
  2. Related entities computed via token overlap, shared country/continent, shared category (ajg_entity_related())
  3. Cross-nav rail listing scopes, desks, and bridges (ajg_render_cross_nav())
  4. Entity-type-specific FAQ of 6-10 Q&A pairs (ajg_entity_faq())
  5. Department views for the 10 flows (pulse/briefs/FAQ/library/encyclopedia/scope-scape/guessing-desk/OPML/print/search)

These are all deterministic functions over the graph. Given the same entity key, they produce the same output. This means every URL is cacheable, testable, and auditable.

5. Editorial standards · what is original vs. quoted

All prose on AJG is original composition. We do not reproduce third-party articles, lyrics, reports, or any other copyrighted material. When we reference primary regulatory texts (Paris Agreement, GDPR, HS Code, Incoterms 2020), we paraphrase into AJG's own voice and always link to the authoritative primary source.

Definitions in our Lexicon (116 terms) are re-compositions from multiple primary sources — we're not plagiarising WCO or ICC texts, we're building a plain-English rendering for practitioners. When a single primary source defines a term (like a regulation's internal definitions), we link out; we do not quote at length.

6. Update cadence · monthly refresh, versioned releases

  • Monthly — comprehensive data refresh runs on the first of each month at 02:00 UTC (admin/monthly-update.php). Regenerates city-master, topic-master, sitemaps, refreshes all counts.
  • Versioned releases — every feature batch is shipped as a versioned patch (v50.x). v50.19 added the universal coverage layer; v50.20 the audit-gap closure; v50.21 the GeoNames 50K-city ingest (planned).
  • Synonym + lexicon expansion — continuous; new trade terms, aliases, and regulatory acronyms added as they're encountered in practice.

7. Audit posture · self-scored against 100-point framework

We self-audit every release against a 100-point Data Intelligence Architecture framework (10 sections × 10 items: inventory, IA/taxonomy, discovery, search-intent, programmatic SEO, mobile UX, data tables, knowledge graph schema, trust, monetization). Current score: 95+/100 as of v50.20.

Current gaps and planned closure dates live in the internal audit dashboard. External researchers can cross-check by comparing our sitemap (sitemap.xml), the entity-view sitemap (sitemap-entity-views.xml), and the rendered JSON-LD on any entity page.

8. Differentiation from commentary

When AJG has original commentary or analysis — as occasionally appears on scope-scape and topic hubs — it is clearly framed with editorial markers. The rest of the site is data + composition + cross-references, presented without editorial interpretation. Practitioners get reference material, not opinion.

For questions, corrections, or enquiries about data sources used on any specific page: enquiry@allfrontierglobal.com.

Totality lens · 32 points to ponder · 16 user POV + 16 developer POV · this institutional hub

User POV — for the practitioner navigating the Methodology institutional hub

Eight dimensions

1 · Possibility

A methodology hub that documents how the platform sources, blends, refreshes, and validates every data field replaces the "trust us" disclaimer with an audit-able surface. The possibility is to win the trust of the most sceptical readers — researchers, journalists, government analysts, expert practitioners — by being more transparent than competitors. Trust earned through methodology disclosure compounds: each new reader who diligence-checks the methodology becomes a higher-conviction user.

2 · Plausibility

Plausibility tracks how honestly the methodology page describes failures. A methodology that claims comprehensive coverage with no gaps is unbelievable; a methodology that names the gaps explicitly ("we do not have ground-truth data for sub-Saharan small-city cost data, we use regional interpolation with confidence flag") is credible. The plausibility ceiling is editorial honesty about limitations.

3 · Probability

On a six-month horizon, methodology-page traffic is small in volume but high in influence. Researchers and analysts cite us based on methodology disclosure; a citation in an academic paper or government report drives recurring traffic that compounds for years. The probability that methodology-investment pays off is high but the payoff is delayed and indirect.

4 · What works

What works is per-data-source provenance. Each major data field carries a methodology entry stating: source, refresh-cadence, verification-method, known-biases, confidence-flag rules. The hub aggregates these entries by domain (cost, climate, safety, etc) so a methodology-curious reader can drill straight to the source they care about.

5 · What doesn't work

What does not work is high-level abstraction. "We use multiple sources blended through proprietary algorithms" tells the reader nothing; it sounds defensive. The hub deliberately avoids that phrasing pattern in favour of source-naming + algorithm-description even when the algorithm is simple, because simple-but-disclosed beats sophisticated-but-opaque for trust.

6 · Common pitfall

A common pitfall is methodology bloat. A 20,000-word methodology page is unreadable; a 2,000-word page is. We chunk methodology by data-domain so each chunk is digestible and the aggregate hub is a navigation surface rather than a wall of text.

7 · Counter-intuitive insight

Counter-intuitively, exposing weaknesses increases trust. Reviewers who arrive sceptical leave with higher conviction when they see the platform name its limitations rather than hide them. The honesty premium is real and persistent — it shows up in conversion rates from methodology-page-readers to engaged-users.

8 · Highest-leverage move

The highest-leverage move is the per-page methodology link. Every entity hub (city, country, vertical, FTA, corridor, tool) carries a "how was this computed" link that goes to the relevant methodology section. Readers who would never visit the methodology hub directly encounter it in context. Per-page methodology links produce roughly 8× the methodology readership of the standalone hub.

Eight user intents

9 · Who gains most

For trust-sensitive readers — researchers citing the platform in academic work, journalists fact-checking platform claims, government analysts incorporating platform data into reports, expert practitioners diligence-checking before professional use, and the methodology-curious sub-group of any audience who want to understand the data before they trust it.

10 · Irreducible essence

They want a clear answer to "where does this number come from and how was it derived". Not marketing copy; engineering disclosure. The schema delivers that at per-data-field granularity so they can drill to the specific source.

11 · Optimal timing

When they are evaluating whether to cite, recommend, or institutionally rely on the platform. Earlier in funnel they read entity hubs; the methodology check is the late-funnel diligence step. Editorial freshness matters because methodology drift (un-disclosed source changes) destroys trust on detection.

12 · Where (sub-areas)

Where they read it: 80 percent desktop because methodology reading is a sit-down task. The mobile design preserves the per-domain navigation but pushes the source-detail tables to expandable cards. Researchers and analysts overwhelmingly diligence-check from desktop.

13 · Why misunderstood

Because trust-by-disclosure is the only durable trust posture in data publishing. Trust-by-assertion ("we are accurate") fails on first mistake. Trust-by-methodology ("here is exactly what we did and where it could be wrong") survives mistakes because the reader was warned about the relevant uncertainty.

14 · Highest-leverage sub-paths

Which methodology-section dominates per audience: source-provenance for journalists, refresh-cadence for institutional users, verification-method for academics, known-biases for sceptical practitioners, confidence-flag-rules for ML-savvy users.

15 · Whose advice to trust

Whose methodology disclosure matters varies by audience. Government analysts care about source-government-data-attribution; academics care about peer-review-friendly references; journalists care about reproducibility. The schema labels each source with the attribution-style appropriate to it.

16 · How to proceed differently

How they use it: enter via a per-page methodology link or direct hub URL, drill to the source they care about, exit either to deeper methodology or back to the entity page they came from. The funnel is intentionally narrow because methodology is a verification surface, not an exploration surface.

Developer POV — for the architect, maintainer, future contributor to this hub

Eight dev dimensions

17 · Data architecture

Data architecture: per-data-field methodology record with source-name + source-URL + refresh-cadence + blending-rule + verification-method + known-biases + confidence-flag-rules + last-updated. The hub aggregates by data-domain. The cron does not refresh methodology directly; methodology updates happen when source-pipelines change, gated through editorial review.

18 · Schema markup

Schema markup: the methodology hub emits as TechArticle with about pointing at the platform itself. Per-domain methodology pages emit as Article with citation children pointing at the source authorities. JSON-LD identifier "ajg:methodology::{domain}".

19 · Internal linking

Internal linking: methodology hub fans down to per-domain methodology pages, across to the entity hubs whose data it describes (every entity hub links back here). Cross-content injector surfaces methodology links contextually when an entity page mentions a data-source.

20 · Page-speed posture

Page-speed posture: methodology pages are text-heavy, no visualisations, no client-side dependencies. Total per-domain methodology page weight under 60 KB compressed. PageSpeed-100-v7 layer applies; the hub is a simple navigation surface.

21 · Mobile UX

Mobile UX: per-domain methodology pages use accordion-style section collapse so readers can navigate to their section of interest without scrolling through the whole page. Source-tables collapse to vertical stacks on mobile.

22 · Accessibility

Accessibility: methodology pages use proper article + section semantics. Source-tables use proper table semantics with scope=col headers. Footnotes are role=doc-footnote with aria-describedby links from the body. Reading-progress indicator on long pages.

23 · SEO saturation

SEO saturation: hub has unique H1 ("How AJG produces its data"), meta-description naming the disclosure-level. TechArticle schema. Per-domain methodology pages each have unique H1 + meta + Article schema. BreadcrumbList. Speakable on summaries.

24 · Extensibility

Extensibility: adding a new data-domain methodology section is a content-only operation; no schema changes needed. Adding a new methodology field across all domains (e.g. "ESG-screening rules") is a schema bump that backfills.

Eight dev intents

25 · Maintainer audience

For the developer maintaining this hub, the most important discipline is keeping methodology current with code reality. When a source-pipeline changes (e.g. we switch from Numbeo-only to Numbeo+Expatistan blend for cost data), the methodology must update in the same ship. We enforce this with a pre-flight check that fails the build if any data-pipeline change lacks a corresponding methodology update.

26 · Architectural commitment

What changes when methodology updates: data/methodology-data.php gains new entries or revises existing ones with version-bumps. The change-log shows what changed and why. Per-page methodology links automatically reach the latest version; old versions are archived but linkable for citation-stability.

27 · Refresh cadence

When the cron runs: weekly at 06:30 UTC on Sundays for the methodology-vs-pipeline drift check. The check compares declared sources in methodology against actual sources in pipeline-config and flags discrepancies for editorial review.

28 · File map

Where files live: data/methodology-data.php (the registry), data/methodology-archive/ (versioned history for citation-stability), includes/methodology-template.php (renderer). Hub at /methodology.php; per-domain pages at /methodology/{domain}/.

29 · Existence rationale

Why version-stable archive: because researchers cite specific methodology versions in academic papers. Breaking those citations by quietly revising methodology destroys the trust-by-disclosure posture. The archive preserves citable versions even after revisions.

30 · Highest-leverage extension

Which renderer: includes/methodology-template.php emits the per-domain methodology + source-tables + version-history + entity-hubs-linking-here rail. Accepts $domain. Echoes directly. Idempotent.

31 · Authoritative sources

Whose responsibility: methodology authoring is shared editorial + dev (dev describes the code reality, editorial polishes for readability). Pipeline-vs-methodology drift detection is automated. Editorial review is the gate for methodology revisions.

32 · Maintenance procedure

How to add a new methodology section: (1) add data-domain to methodology-data.php with the seven-field schema; (2) write per-source entries; (3) verify pipeline-config matches; (4) link from relevant entity hubs; (5) verify hub render. Total: about 4 hours per new domain including writing.

v207.1 cross-Crucible synthesis · Research Methodology

Research Methodology in the cross-Crucible framework

AJG's research methodology has four load-bearing constraints. First, deterministic-PHP composition: all content is generated server-side from typed data files, with zero runtime API calls — the page that loads in your browser at 11:47 AM and the page that loads at 11:48 AM are byte-exact identical (within session-window) because the source-of-truth is on disk, not on a third-party live service. Second, data-anchored every claim: every number, every threshold, every comparative ranking carries a source citation in its host data file (per SO #2). Third, multi-source cross-verification: when sources disagree (Numbeo vs Mercer cost-of-living indices · IMF vs World Bank GDP estimates · Henley vs Arton passport rankings) AJG picks the consensus or surfaces the divergence rather than picking arbitrarily. Fourth, hand-authored prose with seed data: the data is structured (typed PHP arrays with consistent keys); the prose connecting the data is hand-authored to ensure it reads as decision-aid rather than as auto-generated SEO bait. The cross-Crucible angle below maps where each methodology constraint surfaces canonically.

Connect to Crucibles

Knowledge atlas → Where AJG's methodology surfaces deepest — long-form regulatory + technical deep-dives that require multi-source verification + hand-authored connective prose. Knowledge Crucible is methodology's primary execution surface (CBAM mechanics · CSRD reporting · MDR/IVDR · IFRS S1+S2 · GDPR · DSA + DMA). Each entry follows the four-constraint discipline.
Decide atlas → Where AJG's methodology converts to actionable choice — Decide Crucible has the trade-off matrices that thread the data-anchored research into specific recommended decisions. The methodology underpins; Decide is where it concludes.
Cost atlas → Cost Crucible is the most data-quantitative Crucible — Numbeo + Mercer + EIU + UBS + World Bank ICP composite for 197 countries + 1,584 cities. Multi-source-disagreement is highest here (subjective city-ranking variability), so AJG's consensus + divergence-surfacing methodology gets its hardest workout.
Economics atlas → Economics Crucible draws on IMF + World Bank + OECD + national central banks + ratings agencies (Fitch + Moody's + S&P + Coface). The methodology's cross-verification discipline is essential because sovereign-debt sustainability assessments differ markedly across these sources.
Visa atlas → Visa Crucible draws on national-immigration-authority data (USCIS · UKVI · Australian DHA · Canadian IRCC · etc.) + Henley + Arton passport indices + community-source verification (real applicant case-experiences). Methodology requires triangulation here because government-published timelines often differ from real-world processing times.
Business atlas → Business Crucible draws on Doing Business legacy + B-READY successor + Heritage Economic Freedom + Cato/Fraser Economic Freedom + WEF Global Competitiveness. AJG's methodology surfaces both the convergent rankings (Singapore + UAE + Switzerland always top-quintile) and the divergences (Israel ranks differently across these indices for political reasons).
Live atlas → Live Crucible relies on Mercer Quality of Living + OECD Better Life + EIU Liveability + Numbeo + WHO Health Indicators. AJG's methodology applies most heavily here because subjective lifestyle preferences vary and the data-sources weight different dimensions distinctly.
Library atlas → The Library is the operational artefact of the methodology — 7,482+ topics each composed under the four-constraint discipline. Visit the Library to see the methodology in scaled execution; visit Methodology page to understand the underlying discipline.

Related cross-Crucible decision lists

Sources: AJG internal methodology documentation 2024-26 · Server: Nestify Nginx · PHP 8.3 · flat-file deterministic composition (zero APIs at runtime per SO #14) · WTO RTA Database · World Bank ICP + B-READY · IMF DOTS + Article IV · OECD economic outlook + better life index · UNCTAD · Henley + Arton passport indices · Coface + Fitch + Moody's + S&P + EIU country risk · Mercer + Numbeo + WHO QoL composites · WIPO Global Innovation Index · WEF Global Competitiveness · Heritage + Cato/Fraser Economic Freedom