Part 3: Musings from a Data-Science Convention
This is part three in a three-part series by Matt Bell. To start at the beginning, click here.
At lunch, I chatted with a senior geologist from a mid-sized operator in the Permian Basin. I asked him what he thought about the data science “revolution” and how it might affect his work. His eyes twinkled, and he smiled.
“I’m just like any real geologist,” he replied, patiently. “Trying to build models that honor the data and do a decent job pointing my team to the right places to drill and land our wells. I’m sure some of these algorithms are faster and smarter than I am.”
He continued, “But I haven’t seen one yet with the experience to tell which of the patterns it spots are meaningful, and which are meaningless. More than anything, it’s still garbage in, garbage out. If the data’s crap…”
So, what, I wondered aloud, did he make of the kafuffle about artificial intelligence and machine learning replacing human geologists like him? How much longer might his experience save him from being “digitalized”?
He answered, “I suppose that’s what keeps me on my toes. I have to keep on proving that I’m better than the box of electronics on my desk.”
A younger professional joined us who, it turned out, was working for a larger operator on the other side of the basin. She was coming up on three years out of school, having ignored “most people’s advice to go anywhere but the oil industry” in the depths of the downturn.
I asked her the same questions and was intrigued by her answers.
“I got into oil because I wanted to do real geology. I want to work with real data and see the impact of my work on how our wells produce. We still have a lot of work to do to properly understand and model the subsurface.”
Surely, I posited, her generation of geoscientists must already be harnessing data science tools in their day-to-day work?
Another smile. “Of course, we’re trying,” she replied, sounding a tad frustrated. “But I think the expectation that this new generation of leaders and engineers – my generation – is going to work miracles because of data science is setting us up to fail.”
Garbage in, garbage out, I could hear our experienced colleague mumble.
From opposite ends of the career spectrum, they seemed to share a remarkably similar perspective. Each trying to make the best of protozoic data science tools emerging from a primordial algorithmic soup. Each trying to prove their worth, one under extended warranty and the other struggling under the weight of outsized expectations.
I decided to test one more idea before we headed back to our tables for the afternoon session. How did they feel about sharing data between companies?
“We trade data all the time,” he replied. “But you have to have something worth trading. What we really need is more and better data. And that’s hard to come by while we’re stuck in this world of tight budgets and cost control.” She nodded in agreement.
Once again, data science bridged whatever generational and philosophical differences they – or their employers – might harbor. It drove them to understand the inadequacies of their input data and to seek new, lower-cost sources to feed their models and simulations.
As if I needed one, this was yet another reminder that the industry is data-driven. And, while we certainly produce “big” amounts of data as an industry, our particular corner – where subsurface models are built and applied – still needs “bigger and better” data to feed the beast.
We’re busy building communities where like-minded scientists and engineers generate, share, and consume data in new ways.
Although data can certainly become a prized asset, not all data is created equal and an isolationist approach – keeping everything proprietary but consequently not benefiting from others’ data – may soon define the laggards rather than the leaders.
We encourage people to consider the “give to get” approach that Preston mentioned. The more they are willing to give, be that samples to a library collection or data to a shared workspace project, the more actionable information they can get for their limited budget.
As these communities continue to grow, we hope to attract members from a wider and wider demographic. My lunchtime conversations demonstrated that it’s not an old guard vs. new guard situation. Anyone willing to acknowledge the need for more, consistent, relevant data should be intrigued.
Hopefully their employers and, critically, those who invest in such companies will ultimately back and reward those who step up and demonstrate this kind of leadership in the data-driven industry.