Mining data: Goldcorp using analytics to reduce harm.

Author:Kelly, Lindsay

Goldcorp is undertaking a new kind of mining designed to improve the health and safety record across its operations.

Data mining is the process of using analytics to find patterns or trends in data that can help predict outcomes and shape future actions.

Teaming up with financial advising firm Deloitte, Goldcorp undertook a pilot project to examine data gleaned from its Porcupine Gold Mine in Timmins. The goal was to help identify situations in which workers would be more vulnerable to being part of a serious accident.

"What we want to understand is why are some people more likely to put themselves in a situation where they're vulnerable to something bad happening versus other people?" said Paul Farrow, Goldcorp's senior vice-president for people and safety "How can we change our training, our coaching, our leadership development programs to eliminate or, at a minimum, reduce that probability?"

Farrow said 90 per cent of the data examined already exists. Every mine keeps production and safety data, he noted. Analytics help find correlations amongst them.

For example, the data can reveal the effects of the years since the last training course versus the potential to have incidents, or how likely one demographic age group is to have an incident. The company could then look at an age group combined with when it had its last training and see if there's any more detailed correlation.

"What it really showed us was that you can begin to profile an individual or a group of individuals through this data where you know that a propensity for them to have an incident is far higher than another employee set," Farrow said.

From that, the company can start to make changes, whether it's developing training programs designed to address the gap, or changing the daily interaction between supervisors and workers. If you know the profile of the team you're working with, Farrow said, your interactions will be different for different people.

One correlation that immediately stood out is that, as the company gets closer to the end of the quarter, and pressure is mounting to meet production targets, there was a definite increase in the number of incidents that occurred.

Management teams at the mines said it's a parallel they had already intuited. but they didn't have the tangible evidence to back it up.

"When you have some real measurements against these hypotheses, it really makes you think differently" Farrow said.


To continue reading