Property Prediction
& Domaining

Move beyond the manual and subjective - unlock data-driven domaining workflows that better capture your ore body’s complexity

A complete, consistent, and spatially extensive understanding of rock properties and domains is essential for effective mine planning and decision-making, especially in complex ore bodies.

However, many operations still rely on a combination of subjective geological logging and sparse or unevenly distributed property data. This can result in oversimplified ore body models and contribute to suboptimal processing performance and reduced operational efficiency.

Datarock addresses this challenge by developing predictive models that enhance the spatial coverage and accuracy of rock property datasets. We help mining operations move beyond manual, interpretive domain definitions by delivering transparent, robust, and data-driven domaining workflows that better capture ore body complexity.

See the use cases

Datasets

Datarock leverages a wide range of geoscientific and operational datasets – from drill core to processing plant – to enable high-quality prediction and modelling workflows:

Structural and geotechnical data
Hyperspectral and multispectral imaging
Core tray imagery
Mineralogy and petrography
Petrophysical properties
Downhole and surface geophysics

Measure While Drilling (MWD)
Geochemistry
Spectral analysis (FTIR, SWIR, Raman)
Metallurgical testwork
Processing and reconciliation data

Value

Incorporate uncertainty for better modelling using data-driven domains that are flexible and include uncertainty metrics offers a more nuanced alternative to rigid manual boundaries – improving the reliability of downstream models
Optimise spatial sampling strategies by generating smarter, data-informed sampling that targets variability more effectively and reduces unnecessary sample collection
Efficient and adaptive drilling guides drilling programs based on model feedback, ensuring that new data collection is both efficient and strategically focused
Maximise the value of high-cost test work by leveraging broader, lower cost datasets – extending coverage and increasing representativity

Predicting hardness from hyperspectral data

Rock property data is often inconsistently collected, spatially sparse, and influenced by variations in testing techniques over time. By establishing quantitative relationships between existing property measurements and broader geoscience datasets, predictive models can significantly enhance the value, consistency, and coverage of rock property information. These models are delivered through Datarock’s robust NEXUS analytics environment.

Applications

Predict missing or inconsistent data due to changes in test regimes (e.g. evolving assay suites)
Harmonise data collected from different drilling methods (e.g. RC vs diamond drilling)
Extend the reach of expensive or time-consuming test work (e.g. hardness, density, recovery, UCS, petrography, BWI) using high-coverage, cost-effective datasets such as assays, core imagery, and downhole geophysics

Data-driven domaining

The way an ore body is domained fundamentally shapes how it is understood, modelled, and mined. The quality, accuracy, and consistency of these domains—along with their transparency and repeatability—have a direct impact on operational performance and decision-making across the mining value chain.

Datarock specialises in delivering data-driven ore body domains by integrating complex multivariate datasets into transparent, auditable, and robust models. These domains can be generated rapidly and support near real-time modelling workflows for dynamic, data-informed mine planning.

Applications:

Define lithological and alteration zones objectively using multivariate data to minimise bias and enhance model reliability
Classify material based on processing characteristics (e.g. hardness, recovery) to optimise blending, plant efficiency, and throughput
Map zones of similar structural and strength properties to inform ground control strategies and mine design

Let's build something together

Reach out now

Discover other solutions

Prospectivity Modelling

remote sensing, geophysics, geochem, structural.

Click for more

Vectoring & Near Miss

3D, hyperspectral, geochem
Click to know more

Automated Logging

core, chips, ATV
Click to know more