GeaSpirit Evidence Compass

From data to decision

Two different questions

Priority is not Confidence. Confidence is not a probability that a deposit exists. Priority is not a resource estimate. Keeping the two apart is the core of the engine.

Illustrative example

A fictional case — not a real discovery claim — showing what a high-Priority / low-Confidence result means.

This does not mean a deposit has been discovered. It means the evidence is interesting enough to justify investigation, while it is not yet strong enough for high confidence. Positive field/geochemical evidence could raise Confidence; contradictory field evidence could lower Priority and/or Confidence.

What Evidence Compass analyses

Progressive sources are marked as planned integration — not implied to be operational today.

Evidence coverage — more data is not automatically better

A source should increase confidence only when it provides relevant, sufficiently independent and useful evidence. Missing evidence must not automatically mean poor geology.

Evidence Compass aims to distinguish availability, quality, independence, relevance and consistency — not just count sources.

Counter-evidence — not only reasons to say yes

The engine is deliberately designed to look for information that could weaken a hypothesis, so it does not simply accumulate positive signals.

Conclusion: “Interesting, but not yet strongly supported.”

Next best action

Evidence Compass proposes the most rational next step — not a fixed sequence. It never equates “acquisition due diligence” with BUY, or “drill target review” with DRILL HERE. Human expert review is always required.

A second life for historical mining material

Historical tailings and mine waste may contain remaining valuable materials — but material may already be processed, grades can be low or heterogeneous, recovery can be difficult, and environmental, legal and economic viability must be demonstrated. Evidence Compass only helps identify which cases may justify specialised study.

Deposit-specific reasoning

The engine does not use the same evidence for every mineral system. This is an illustrative, provisional domain model — not the internal weights.

ILLUSTRATIVE / PROVISIONAL DOMAIN MODEL — actual internal weights are not published.

Known asset scale — separate from Priority

A globally famous world-class deposit may carry Known Asset Scale = WORLD CLASS while Evidence Compass Confidence is LOW, if the evidence available to the engine itself is incomplete. Evidence Compass does not pretend to “rediscover” known facts because they are famous — known size and evidence-based screening are different concepts.

How it reasons (conceptual)

# CONCEPTUAL PSEUDOCODE — NOT THE PROPRIETARY ENGINE
asset = collect_available_evidence()

priority   = evaluate_interest(geology, independent_signals,
                               strategic_context, resource_context)
confidence = evaluate_support(evidence_quality, independence,
                               completeness, contradictions)

result = {
  "priority":   priority,      # should we care?
  "confidence": confidence,    # how sure are we?
  "why":        strongest_evidence,
  "against":    counter_evidence,
  "missing":    critical_unknowns,
  "next_action": recommended_investigation
}

No real code, numerical weights or thresholds are shown. The engine is private.

Independent evidence beats duplicated evidence

Two datasets are not automatically two independent confirmations. If several layers share the same underlying information, Evidence Compass tries not to count it twice — independent evidence is more valuable than duplicated evidence.

A system that must keep learning

The volume of geological and Earth-observation data grows continuously, so the screening architecture cannot be static. Evidence Compass evolves through a controlled cycle: new data → curation → quality/provenance → evidence coverage → model update → validation → new version (v1 · v2 · v2.x) — controlled evolution, not uncontrolled self-modification.

High reliability has to be demonstrated, not assumed

Software running successfully does not prove geological reliability. GeaSpirit challenges the engine against known cases, negative cases, contradictory cases, real-world evidence and version-to-version benchmarks — and, where possible, future external review. Evidence Compass is currently in validation phase. Its geological reliability is not presented as independently validated or accuracy-certified unless supported by external evidence.

What Evidence Compass is — and is not

Three-layer GeaSpirit architecture

4. Subsurface Intelligence CONCEPT

Today GeaSpirit works mainly as a large-scale screening and prioritisation layer over a global universe of assets. Subsurface Intelligence is the planned next generation: once an asset has been shortlisted, it will bring together the available public and (with permission) private technical evidence — geology, drillholes, assays, geophysics — into a shared spatial view, and feed that evidence back into Evidence Compass.

Planned modules

Secondary resource intelligence in 3D

Planned extension for historical tailings and mine waste: surface extent, morphology and DEM, historical processing, sampling, mineralogy and remaining critical-raw-material evidence combined into a 3D volume with uncertainty, to prioritise which secondary resources deserve physical characterisation. No resource or economic claim is made without appropriate physical and technical evidence.

Illustrative example — from global screening to a subsurface target

The figures and targets below are synthetic and illustrative only — not real resource estimates, not real projects.

Public data + private customer data

GeaSpirit uses open evidence extensively today. Subsurface Intelligence is intended to let authorised customers combine it with their own private drillholes, assays, mapping, geophysics and interpretations. Subsurface Intelligence will be designed to keep authorised customer data private and isolated from public datasets. Private customer information will not be made public automatically, while GeaSpirit's proprietary scoring mechanics will remain private.

Designed to work with existing geological workflows

GeaSpirit is not intended to force companies to replace their existing geological-modelling or exploration software. The planned objective is interoperability through open, non-proprietary formats (e.g. CSV, GeoJSON, GeoTIFF, LAS where applicable, DXF, Shapefile, VTK, standard drillhole tables) so it can sit alongside GIS, geophysical, database and modelling tools.

No official integration with any third-party product is claimed. Interoperability categories are described conceptually and will be implemented and tested progressively.

GeaSpirit today vs Subsurface (planned)

Future technology under evaluation

Possible future building blocks include 3D visualisation, implicit geological modelling, geospatial and volumetric processing, drillhole processing, geophysical grids, uncertainty modelling, customer workspaces and API integration. Open-source components such as GemPy, PyVista and VTK are under evaluation only — not part of any production system — and a licensing / freedom-to-operate review will precede any commercial use.

Development roadmap — all stages planned

Development will proceed progressively through technical validation, industrial testing and customer-supported pilots. Every stage above is planned unless working, tested code proves otherwise.

Explore GeaSpirit