Research & intelligence methodology

Scientific methodology. Continental-scale intelligence.

The Convergence Score is built on peer-reviewed geoscience — validated spatial algorithms, multi-source data fusion and auditable machine learning — producing defensible, investment-grade intelligence.

Read the methodology ↓Download white paper

Research architecture

Three methodological pillars.

Each framework is independently auditable, versioned and reproducible.

Multi-source convergence

The Convergence Score is the weighted sum of four independent intelligence channels, each normalised and summed into a 0–100 investability metric for every target.

Every update to an input layer triggers an automatic recalculation with a new version stamp — a full audit trail of how each score evolved.

// Convergence Score
score = drone   // 0–40
      + archive // 0–30
      + scout   // 0–20
      + geology // 0–10

if score >= 70 → CERTIFIED_TARGET

Validated spatial intelligence

Scout reports are validated before they count. Spatial significance uses DBSCAN — the standard for finding arbitrarily shaped clusters in geospatial data.

Reports failing GPS plausibility, duplicate detection (image hash + coordinate proximity) or HMAC session checks are quarantined.

// DBSCAN configuration
algorithm:       DBSCAN
epsilon:         5 km
min_samples:     3 validated reports
distance_metric: haversine
index:           BallTree

Deposit Potential Index

A multi-criteria framework that normalises diverse historical data into the 0–30 archive contribution. DPI is re-computed and timestamped whenever new documents are ingested for a target.

Host rock typeLithological affinity and mineralisation potential

Production recordsHistorical output volumes and grades

Evidence densityMINM-indexed documents referencing the target

Mine lifespanDuration of recorded extraction

Structural controlsDistance to fault and fold axes

Current ASM activityActive scout clusters within 10 km

FORGE Africa Institute · Research partner

Methodology co-developed and peer reviewed.

AfriXplore co-develops its methodology with FORGE Africa Institute, a pan-African research network in geological intelligence and mineral resource assessment.

Archive research is conducted with local geological authorities across sub-Saharan Africa. Every methodology update undergoes external peer review through FORGE Africa's academic network.

About FORGE Africa Institute →

Research collaboration

Joint work on archive digitisation, DPI frameworks and score validation.

Academic network

Geoscientists across the region for peer review and field validation.

Data sovereignty

Direct collaboration with national geological authorities.

Institutional credibility

External validation to investment-grade standards.

Geophysical methodology

Anomaly detection at multiple depths.

Three independent datasets per Geoswarm survey, each processed with industry best practice.

Magnetics

Geoswarm

TMITotal magnetic intensity — primary anomaly map

RTPReduction to the pole

1VDFirst vertical derivative — shallow features

Euler deconvolutionSource depth estimation

Gravity

Geoswarm

Bouguer anomalyTerrain-corrected complete Bouguer gravity

Horizontal gradientEdge detection of density contacts

Free-air anomalyRegional isostasy

Upward continuationRegional field separation

Hyperspectral

Geoswarm

MNF transformMinimum noise fraction

SAMSpectral angle mapper — mineral mapping

MTMFMixture-tuned matched filtering — abundance

Feature extractionSWIR absorption features

WHITE PAPER · 40 PP

The Convergence Score Methodology

AFRIXPLORE RESEARCH · 2026

White paper

The full scoring framework, in 40 pages.

DBSCAN parameterisation, DPI weights and the Certified Target process — with worked examples from West African gold-belt targets.

Questions about our methodology?

Talk to our research team about validation, peer review or partnerships.

Contact research team →