Mineral exploration data platform: drilling, geochemistry, geophysics and GIS in one place
A mineral exploration data platform brings decades of geological reports, drill holes, assays, geochemical and geophysical surveys and satellite imagery into one environment with one coordinate system, so a geologist can query them together instead of reconciling spreadsheets, PDFs and GIS layers by hand. Mirage builds that environment around the datasets an exploration team already holds and applies machine learning where it helps rank targets; the decision stays with the geoscientists.
The problem
Exploration information is spread across decades of geological reports, maps, drilling campaigns, geochemistry, geophysics, satellite imagery and databases, often in incompatible formats built at different times by different teams, in two or three coordinate systems.
The challenge is not having more data. It is making the datasets that already exist usable together: the 1998 trench samples with the 2021 drill holes, the airborne magnetics with the soil geochemistry, the scanned report with the GIS layer.
What we deploy
A unified exploration environment that combines these geological datasets in one coordinate system and applies machine learning where it is useful: to identify patterns, analyse anomalies and help prioritise exploration targets. The geoscience team queries it, draws on it and decides from it; it does not drill anything on its own.
Datasets, outputs, systems
| Reads | Produces | Connected through |
|---|---|---|
| Drill hole collars, surveys, lithology and assay intervals | Validated holes in one database, with desurveyed traces and composited intervals; overlaps, gaps and unit mismatches flagged | Drilling database (acQuire, Micromine, Leapfrog, Excel or CSV exports) |
| Soil, stream, rock-chip and trench geochemistry | Normalised element grids and anomaly maps, with detection limits and laboratory batches kept | Laboratory certificates (CSV, PDF), GIS layers |
| Ground and airborne geophysics: magnetics, gravity, IP, EM | Gridded and filtered layers in the common coordinate system, overlaid with geology and geochemistry | Survey deliverables (grids, line data), GIS |
| Historical reports, maps and sections, scanned or digital | Text and figures indexed and georeferenced; intervals, samples and conclusions extracted with their page reference | PDF and image archives, read as documents |
| Satellite and remote sensing imagery | Alteration and structure layers, aligned with the ground datasets | Public and licensed imagery, GIS |
| Tenements, access, infrastructure, environmental constraints | Ranked targets filtered by what can actually be drilled | GIS and cadastral layers |
Outputs stay in the tools the team uses: GIS layers, 3D models and drill databases are written back, and the ranked targets come with the evidence behind each rank.
How it works
Mirage engineers work with the geoscience team to inventory and normalise the existing datasets, historical reports, drilling logs, geochemical and geophysical surveys, satellite and GIS layers, into a single environment with one coordinate system and one set of units. Machine learning is applied where it adds value: flagging geochemical or geophysical anomalies, correlating patterns across datasets that were previously reviewed in isolation, and helping rank areas for further fieldwork.
Most of the work is the normalisation. A model trained on badly joined data ranks the join errors, which is why the inventory and the validation of every dataset come before any model is run, and why each ranked target carries the samples, holes and layers that produced it.
What the machine learning does, and what it does not claim
| Task | What the system does | What stays with the geoscience team |
|---|---|---|
| Anomaly detection | Flags samples, intervals and grid cells that stand out against their neighbours and against the dataset's own distribution | Deciding whether an anomaly is geology, contamination or a laboratory artefact |
| Pattern correlation | Finds where geochemistry, geophysics, structure and alteration coincide, across datasets reviewed separately until now | Interpreting the coincidence in the deposit model |
| Target ranking | Orders areas by the evidence that supports them and by what can be drilled, with the evidence listed | Choosing the targets, the drill plan and the budget |
| Historical data | Extracts intervals, samples and conclusions from scanned reports, with page references | Judging the reliability of a 1980s report |
| Prediction | Does not claim to find minerals or to predict grades | All of it |
Operational interface
Decision-support infrastructure for exploration teams, not a claim that the system finds minerals on its own.
- Geological maps and satellite imagery, one coordinate system
- Drill holes and geochemical layers, queryable together
- Geophysical data overlays
- 3D subsurface visualisation
- Anomalies and ranked exploration targets, each with its evidence
- Historical reports indexed and georeferenced
How it is deployed
Forward-deployed: engineers with your geologists, from the data inventory to the first ranked targets. The proof of concept usually takes one project area with its drill database, assays and one geophysical survey, so the team sees its own data joined before any commitment.
- First reply
- Within two business hours, from an engineer.
- Proof of concept
- On your own data (a telematics export, a transaction log, weighbridge tickets, dispatch records), one week after the first meeting.
- Pilot
- Fixed price, on one site or one fleet, with Mirage engineers deployed with your team. You can stop after it.
- Production
- The first workflow runs in 4 to 6 weeks. A system that spans several sites or a lot of field equipment is scoped at the first meeting and can take longer.
- After go-live
- Mirage operates, monitors and improves the system in production; your team keeps the decision on every exception. More than 200 systems have been connected so far.
- Price
- An annual subscription priced by volume after the pilot, from $10,000 to launch and $1,000 a month, quoted in the first conversation.
Mirage has not published an exploration deployment; its named mining references are Moroccan operators running fleet, fuel and control-room systems, and mining is offered internationally. An exploration engagement starts with the inventory of what the team already holds.
FAQ
See what Mirage can run in your operations
Get a custom automation plan
Tell us where your operations lose time. We send back a written plan: which workflows are worth automating, what a deployment looks like on your systems, and the shortest path to production. No commitment.
Prefer to talk it through? Book a call