learn

Architecture

How DGeo Green structures brand data using the 3-layer Essential-Infological-Datalogical model

Overview

DGeo Green structures every brand entry using a 3-layer architecture inspired by MDA (Model-Driven Architecture) and REA (Resource-Event-Agent) ontology. This ensures brand data is meaningful to humans, machines, and AI crawlers simultaneously.

          text
        
┌─────────────────────────────────────────────────┐
│  Essential Layer (CIM)                          │
│  WHY this brand is indexed                      │
│  Intent, mission, compliance category           │
├─────────────────────────────────────────────────┤
│  Infological Layer (PIM)                        │
│  HOW brands relate                              │
│  REA graph, cross-references, prerequisites     │
├─────────────────────────────────────────────────┤
│  Datalogical Layer (PSM)                        │
│  CONCRETE data                                  │
│  Contacts, certifications, FAQ, articles        │
└─────────────────────────────────────────────────┘

Essential Layer (CIM)

The Computation Independent Model — human-readable brand intent, completely independent of implementation.

For each brand, the Essential Layer captures:

Field

Purpose

Example

`title`	Brand name	Keripik Keladi Asin Pelangi
`description`	One-line value proposition	Halal-certified Papuan taro chips with green energy production
`audience`	Who this page serves	consumer, distributor, developer
`purpose`	Page intent	learn, purchase, partner
`priority`	Visibility weight	critical, high, medium
`requires`	Prerequisite knowledge	certification/halal, certification/green-energy

Example

          yaml
        
title: Keripik Keladi Asin Pelangi
description: >
  Halal-certified Papuan taro chips produced with green energy
  and AI-powered commerce from Sorong, Papua Barat
audience: consumer
purpose: learn
priority: high
requires:
  - certification/halal
  - certification/green-energy
  - certification/ai-commerce

Infological Layer (PIM)

The Platform Independent Model — an REA ontology graph that structures brands into formal relationships.

REA Mapping for Brand Indexing

REA Concept

Brand Index Mapping

Resource	Products (keripik, beras), services (homestay), certifications
Event	Production, certification audit, purchase, booking
Agent	Producer (UMKM), certifier (MUI/BPJPH), consumer, distributor

Relationship Types

The content graph builds typed edges between brand pages:

prerequisite — "Read about halal certification before this brand page"

related — "This brand is in the same region as another"

next/prev — Sequential navigation within a group

cross-reference — Inline links between pages (auto-detected)

Graph Visualization

          text
        
certification/halal ──prerequisite──▶ brands/keripik-keladi
certification/halal ──prerequisite──▶ brands/beras-marolis
certification/halal ──prerequisite──▶ brands/homestay-farasman
certification/halal ──prerequisite──▶ brands/nut-tonton

certification/green-energy ──prerequisite──▶ brands/keripik-keladi
certification/green-energy ──prerequisite──▶ brands/homestay-farasman
certification/green-energy ──prerequisite──▶ brands/nut-tonton

brands/homestay-farasman ──related──▶ brands/nut-tonton
  (same region: Raja Ampat, same category: halal tourism)

Datalogical Layer (PSM)

The Platform Specific Model — the concrete, rendered data for each brand:

Contact details (name, phone, email, address)

Certification IDs and status

Product images and gallery

FAQ (structured Q&A pairs)

Articles (long-form content)

Testimonials (verified reviews)

Location coordinates and regional data

Example Data Structure

          json
        
},
  "features": {
    "green_energy": true,
    "ai_commerce": true,
    "halal_certified": true
  }
}

GEO Advantage

By structuring every brand entry across all three layers, DGeo Green provides:

AI-readable metadata — Essential Layer fields are machine-parseable

Knowledge graph — Infological Layer creates navigable relationships

Structured data — Datalogical Layer provides concrete facts for extraction

llms.txt compatibility — All three layers are exposed to AI crawlers

This means when an AI agent asks "What halal-certified food products are available in Papua Barat?", the graph can answer it directly — not through keyword matching, but through typed relationships.

Methodology

Keripik Keladi Asin Pelangi