Datarock unearths value by bridging the gap between data collection and data-driven decision making
Transitioning to a renewable economy will require a new mining industry that can extract significantly more metal from the ground to meet future demand, while operating in accordance with stricter environmental, social, and governance responsibilities. However, this is compounded by a myriad of challenges. Exploration success rates are dwindling, existing deposits are becoming deeper and more arduous to mine, and ore grades are declining.
Mining companies are collecting more data than ever, yet they struggle to maximise value to meet the greater demands placed on them. The gulf between data collection and data-driven decision-making can be significant.
In recent years, the mining industry has realised that machine learning (ML) is key to gaining the full value from its data. This transition will be critical to the massive optimisation needed to meet the world’s growing appetite for critical metals.
Datarock is one company on a mission to help miners explore and extract resources more efficiently by maximising the value of their geoscience data.
Their team of experts build and deliver solutions for mining companies that essentially ingest data and export answers by leveraging ML and cloud computing. The company is currently building out its AI geoscience platform, which will allow miners to efficiently extract information from geoscience data.
Founded in 2018, Datarock has grown to over 50 staff and garnered significant investment from IMDEX –the Australian mining technology giant.
A Geology First Company
Datarock considers itself an Ore Body Knowledge (OBK) company that’s focused on developing innovative products, which allow mining companies to increase their understanding of their deposits.
At the heart of Datarock’s geology-first mindset is their Applied Science division. This revered team of geologists, data scientists, and ML engineers represents the largest of its kind globally, one that has worked on over 200 data science projects for the world’s largest mining companies. The technical team solves some of the hardest problems in mining and exploration, building solutions and driving R&D and product development for Datarock.
Datarock Chief Geoscientist and Technologist Brenton Crawford explains how Datarock develops their technology and products.
“One of the things we identified early is that you can’t afford to make a mistake and develop a product that doesn’t ultimately deliver significant improvements and efficiencies for a customer. The work performed by our Applied Science division ensures technology development is at the forefront of scientific advancement and aligns to the practical needs of the mining sector,” he said.
Datarock has deployed a large number of ML models inside global mining companies. These models ingest geoscience data and export answers that allow miners and explorers to make data-driven decisions. The company has utilised this knowledge and experience to develop an AI platform that enables these types of workflows to be simplified, scaled, and used day in, day out.
A platform to augment and automate geological logging
Datarock’s first software offering is a SaaS solution that allows customers to upload images of drill cores to an ML platform that automatically interprets the imagery and generates high-value datasets.
Crawford explains how this solution addresses some of the long-standing inefficiencies from data collected in the coreshed.
“Traditionally, geotechnical and geological data are manually logged by geologists and geotechnical engineers.
“Manually logged datasets are plagued by inconsistency, inefficiency, and lack of an audit trail. By using computer vision (machine learning) to interpret the core imagery, we can generate similar types of data that manual logging is capable of, but without many of the pitfalls common in traditional methods”.
Datarock created its innovative ML platform to generate consistent, high-resolution and quantitative datasets from drill core imagery. This allows mining companies to augment and automate core logging practices to generate geological and geotechnical data at scale, with speed and transparency, which is impossible using traditional methods.
Crawford added that core photos are rich in geological information, and one of the most frequently collected datasets across an entire deposit, however they’re highly underutilised.
“By extracting data in this way we can really add significant value,” he said.
“This value typically presents itself to clients in three key ways. Having data with a digital audit trail allows high value decisions to be made in a transparent way. An example might be if an engineer needs to create a mine design, but the drill core data they are using has no audit trail, they may be more conservative and create a sub-optimal design. But if they can see exactly how that data was created transparently, then that might translate to significant savings for a mining client.”
Australia currently faces a geologist shortage, and operations are being asked to collect more data than ever before with fewer people available. Datarock staff can help free up mine personnel, allowing them to work on higher value tasks, leaving the head office to manage elements of data collection remotely.
“A lot of the more expensive and impactful datasets an operation collects are commonly not collected continuously due to the high cost,” said Crawford.
“Core photography is typically done on all diamond drilling, which means that if we are generating data from these images, we have the opportunity to create them across the entire deposit. We often refer to these high value full coverage datasets as foundational data.”
Datarock’s software continues to evolve rapidly, collecting and interpreting new types of geoscience data, and its next addition will be capable of analysing rock chip imagery in a similar way to drill core.
For more information on how Datarock’s Platform and Applied Science can help your business, email [email protected] or visit https://datarock.com.au/
