The youth of today. No longer using anarchy to upset the established order; now, they're turning to data.
Connor Tann is one such data revolutionary. And he’s not alone. Some 800 more digital devotees have rallied to his call, joining an online army of BP scientists and engineers growing their data science skills in their spare time.
“There’s a massive race on in the data science field,” says petrophysicist Tann. “And BP needs to be a leader in the pack.”
In internet chatrooms and BP’s canteens, his army furthers its goal: widespread disruption of the digital kind. And their mission is showing signs of success.
BP's Upstream business is transforming as traditional ways of working are sped up and convention is turned on its head.
One weapon of choice is machine learning. It's where computers are trained to spot relationships and make predictions as scale. And it is proving pivotal in tackling one of the industry's biggest challenges: where to find oil.
Advances in imaging technology have made huge improvements in the visualization of layers of rock deep beneath the earth’s surface but it still leaves a lot of uncertainty about where pockets of oil and gas might be found. This is where the ability to interrogate vast amounts of data can reduce the speculation around where to drill expensive wells and how to best maximize production.
“What we’re trying to do all the time is accurately describe the full range of possibilities in a reservoir to help BP make the best possible financial decisions,” says Tann who has a masters in experimental and theoretical physics from the University of Cambridge.
“To help us reduce uncertainty - machine learning allows us to run tens of thousands of possible subsurface scenarios much more rapidly than before − essentially, helping us look for a very small needle amongst lots of haystacks."
Machine learning is also revolutionizing the nature of traditional tasks.
“As a petrophysicist, I’d typically have to interpret each of my wells, one by one,” says Tann.
“Machine learning allows you to do that in bulk. You can teach a machine to learn by example from a fraction of your existing data, before unleashing it on the rest.
“Basically, it is super-charging the ability to understand each oil well, saving the business a lot of time and money.”
On latest estimates, Tann's machine learning application reduces the time it takes to interpret the data from each well by 80%, meaning what once took months now takes days.
That leaves talented young data scientists like Tann time to do the interesting work. “We don’t spend so much time doing manual data drudgery,” he says.
“Machines can do loads of things far better than humans and dealing with huge amounts of data is one. It's very hard for a human to look at that enormous data set because, quite simply, there is no way you can put it on one screen,” says Tann.
But the machines aren’t taking over.
“There are lots of things humans can do a lot better than a computer − things involving emotional intelligence or communicating, or explaining the reasoning behind business decisions."
It is Tann's huge intellect and his 'softer' skills that brought him to the attention of BP’s team at the top. A year ago, Upstream chief executive Bernard Looney selected Tann as his ‘reverse mentor’.
Each month, the pair meet to discuss and see demonstrations of the latest trends in data science. It’s an area about which Looney is passionate - ensuring BP is fit for the digital age with the right behaviours, mindset and culture to match.
Looney says: "I'm passionate about our modernization and transformation agenda. And this mentorship continues to inspire me to provide a place for people like Connor to have a great career.
"And I feel and enormous obligation to ensure that happens. Connor's intellect is massive, yet he is very down to earth with no airs or graces."
And, as well as having a direct line to a senior BP exec, Tann has unlimited access to one of the biggest supercomputers in the world.
“I can log on to the CHPC, the center for high-performance computing in Houston, where they’ve ring-fenced quite a few machines to data scientists like me. It’s called the data science sandbox. It’s a playpen for building models.”
It may sound like child’s play, but this is serious business for Tann and other digital experts like him.
“It’s just an amazing asset that we have at our disposal,” he says.
Combining that asset with minds like Tann’s is speeding the revolution of BP’s upstream.