Data mining: How digging through big data can turn up new mineral deposits

A team of mineralogists is using the tools of network analysis to understand connections between the Earth’s minerals

A lump of parisite-(La), a new mineral predicted by big data techniques and found in north-eastern Brazil.
A lump of parisite-(La), a new mineral predicted by big data techniques and found in north-eastern Brazil.
Luiz Menezes

Mineralogy and mining are two things that often go hand-in-hand. But scientists at the Carnegie Institution for Science are mining something different – big data, and it may have a significant impact on how minerals are studied and discovered.

There are at least 5200 different known minerals (naturally occurring chemical compounds not formed by biological sources), which have been found and catalogued in hundreds of thousands of different locations around the worl. This gives mineralogists millions of data points to work with, but extracting meaning from this data is often difficult.

In a new paper published in American Mineralogist (PDF), a team of scientists have used network theory to gain new insight into the distribution and changes of copper and chromium deposits over time. This approach could lead to the prediction of new minerals or the discovery of new mineral deposits.

Network theory is a way of analysing complex connections and interactions between different objects. This method is often used to analyse the spread of disease, the structure of the internet or biological systems. The Carnegie team used this approach to assign each known mineral as a “node” and each location where two minerals were found together as a connection between these nodes. This generates a 3D network, from which trends and clusters can be identified, pointing scientists towards locations where precious mineral deposits might be uncovered.

“The quest for new mineral deposits is incessant, but until recently mineral discovery has been more a matter of luck than scientific prediction,” says Dr. Shaunna Morrison, one of the lead authors. “All that may change thanks to big data.”

The technique can also be used to predict the existence of previously unknown minerals. The team have predicted 145 unknown carbon-containing minerals, of which 10 have now been discovered.

Network theory can also be used to understand how the Earth’s geology has changed over time, and how it has been affected by living organisms. The authors describe how the distribution of copper-containing minerals has changed as the Earth’s atmosphere became rich in oxygen. It is hoped that similar analysis can be applied to planets such as Mars, offering a glimpse into the geological – and maybe even the biological – history of the red planet.

Joel Hooper is a senior research fellow at Monash University, in Melbourne, Australia.
Latest Stories
MoreMore Articles