Leveraging Big Data for Mineral Exploration
Novel ways of using big data to discover new mine exploration sites.
OZ Minerals, Australia
Among the many advantages of open innovation is that you can bring together a lot of very smart people to work on a problem for you, potentially hastening the arrival of a solution.
When mining company OZ Minerals wanted help to discover new exploration targets at a site near its Prominent Hill copper-gold mine in South Australia it reached out to the crowd with the launch of the Explorer Challenge.
Mining the Data
Like other mining companies, OZ Minerals pores through a lot of data to pinpoint the most promising drill targets. The problem with this according to its CEO Andrew Coles is that the company only has a small team of people who explore for new mineral resources, but they work on multiple projects and only have “limited time and bandwidth to actually look at data.” Hence the need to bring in a bigger pool of people from around the world.
“This Explorer Challenge will hopefully give us access to the world's smart people,” he continued. “Many of whom will look at data in a very different way to what we look at data. So we are very hopeful that we are going to get very different perspectives using multi-layered data sets to help us in our interpretations.”
The open innovation competition ran for three months and concluded in May 2019. During that time, more than 1,000 participants from 62 countries registered to take part. To help them develop their novel big data approaches they were given access to more than six terabytes of public and private exploration data, such as geochemistry, drill hole and seismic surveys.
The top three solutions shared a prize pot of 800,000 Australian dollars. In first place was Team Guru for an approach that included interpretable machine learning models for mineral exploration using geochemistry, geophysics and surface geology.
DeepSightX was in second place for an artificial intelligence model to locate promising exploration targets while third place was awarded to a team from Cyence, a platform for the economic modeling of cyber risk for the insurance industry. They used data science to select the ten best examples of candidate point estimates.
Several other prizes were awarded including the Student Team prize and the Genius Prize.
Following the completion of the competition, the winning solutions are going to be tested in the field. The top targets are scheduled to be drilled by the close of 2019.
Speeding Up the Discovery of Drill Sites
Turning to external smarts has the potential to drastically speed up the lengthy and intensive mineral exploration process.
In commenting on the success of the challenge Andrew Cole was clear about the benefits of an open innovation approach. He said:
“The innovators who participated in the Explorer Challenge have provided approaches to mineral exploration that we never would have imagined internally, including ways to fuse datasets together, combining multiple layers of information, and making predictions based on the extensive datasets.”
Next Story »