PETRA® Data Science enables resource companies to turn data into real-time predictions and optimisations. We help our clients lead the world in preventing downtime with real-time machine learning algorithms. Proprietary big data integration services enable our clients to break down data silos and optimise across the whole operation (e.g. mine to mill, pit to port etc).

PETRA brings together a transdisciplinary team of PhD qualified engineers, mathematicians and programmers whose sole purpose is to engineer data science solutions for the resources industry. The diversity of our team enables us to think and act beyond traditional silos of domain expertise.

We are proud of the years of site experience and industry expertise that enables us to find hidden value in real mine data. Because we are resources professionals, we won’t take up valuable site personnel time trying to understand the data.

Our FORESTALL® real-time alert algorithms work with any process control/visualisation system, including; IoT end node processing/control (DCS), corporate data visualisation platforms (QuickView, Tableau and SharePoint), and platform agnostic cloud processing. Our 2016 Mining Magazine Award finalist PRODFINDER® data integration solution integrates 3D spatial data with time series data (PI, SCADA) to enable big data mine to mill optimization


Platform agnostic and industry proven algorithms prevent downtime

Our proprietary data integration and prediction services are designed to support your mining operations with insights, algorithms and predictions within eight weeks. Our 2017 Austmine Innovation Award finalist FORESTALL® real-time alert algorithms are platform agnostic and work with any process control/visualisation system (e.g. DCS and PLCs), corporate data visualisation platforms (QlikView, Tableau and Sharepoint), and cloud processing (e.g. GE Predix, Microsoft Azure and Amazon Web Services).

Value chain optimisation

Every mine can collect data, but making it usable and integration 3D spatial data with time series process data can be an artform. For example, our 2016 Mining Magazine Award finalist PRODFINDER® data integration enables big data mine-to-mill type projects to be completed for a fraction of the cost and time of traditional methods.

Harnessing the power of machine learning for exploration

Machine learning AI has emerged as a trend that is helping all manner of operations to make the most out of their data. Our advanced machine learning algorithms are trained to automate code matching across thousands of fields. Tests show that tasks which used to take three to five days, can now be completed in 15 minutes!

Bespoke data science applications

Our creative team of mathematicians, engineers and programmers love new challenges and we regularly work inside resource companies to solve a specific challenge. For example, our FRAGxTM algorithms quantify fragmentation using lidar point cloud data, whilst at the opposite end of the rock scale, LABYRINTH® detects microdamage using x-ray microtomography data.