Interview: Professor Mike McGuirk on How Brandwatch For Students is Used in His Classroom
By Olivia SwainSep 6
Published August 23rd 2019
As a woman with a numbers background, I have always been curious about data science.
Since I’m interning at Brandwatch, I took it upon myself to seek out the data science team to see what I could learn and I managed to grab Hamish Morgan, VP of Data Science, for an interview.
Hamish came from an technical background, studying for four years towards a PhD in Natural Language Processing. He quickly found his calling in a managerial role, and now seems to be very happy talking to me about all things data science.
So, what exactly is the data science team and what do they do?
The team is 15 people strong, including a couple of summer interns like myself, and everyone, regardless of their background or experience, is referred to as a data scientist.
I asked Hamish to describe the team in a sentence.
“We are developing new products and features that empower our customers to become data scientists and find the insights that deliver real value.”
But there’s a lot more to it than that.
“Our customers are trying to ask and answer research questions on social data, but most of them don’t have a statistics background. We’re trying to build the products that support them in asking data science questions.”
His team are working on solutions that can help a diverse set of customers generate valuable insights. It’s the variety of use cases that our customers have that keeps the team on their toes. “A normal data scientist solves the problem once. We have to solve the problem for all possibilities and productize that,” he says.
The data science team has existed for nearly three years in Brandwatch, and was originally formed of people from the engineering team who had the relevant skills.
“Over that time we’ve grown and learned a lot about how to do this,” Hamish says. “The team is much more diverse in terms of its backgrounds; much less engineering and much more other things.”
The team is ever changing, “just like Brandwatch.”
I queried Hamish on the variation of the team’s backgrounds, and how the team works together given the range of different sets of expertise.
“Within the data science team I try to make sure we’ve got a broad range of different backgrounds and different experiences,” he says.
From engineering to psychology, there have been a whole variety of paths that led people to data science at Brandwatch.
“Cognitively diverse teams are more creative, better at problem solving, and generally work faster. I think departmental diversity is absolutely key to creativity.”
As the merger and new integrated product is on everyone’s mind, I felt I had to ask how that has affected the running of the team. Crimson Hexagon’s equivalent team have combined with the existing Brandwatch data scientists to form an even more diverse team that stretches across the globe. “We’ve found that we had very different, but very complementary skills,” Hamish explained. “Since the merger we’ve been collaborating and working really well together. We’re a much stronger team than either of us were independently.”
Still curious about how the team themselves define data science, I asked Hamish about the use of the word ‘science’ and whether that can lead to misconceptions.
“Scientific method is an important part of what we do, but it does also create a sense of aimlessness. We’re much more applied, we’re trying to solve customers’ problems by looking at what’s available and what technology we can apply.”
According to Hamish, “pointy headed academics sitting in ivory towers shouting Eureka” is a common perception of the data science team internally. As the team aren’t as visible as they could be, he understands where this comes from.
“The reason we’re quite quiet is that our projects take a long time and frequently fail. We fail a lot. And that’s a good thing, if you’re not failing then you’re not trying hard enough. You’re not taking risks. For every ‘Iris’ that sees the light and is super impactful, there are going to be several projects that died quietly.”
Despite the team’s tendency to be quiet, Hamish acknowledges the growing need for more visibility for the team both in the company and externally.
They’ve recently introduced a role that sits across the commercial and engineering departments, and they’ve even hosted ‘fireside chats’ that allow anyone in the company to ask questions of the team.
And there’s more to be done.
“An important secondary aspect of the team is that we should be trying not just to build the products but actually provide leadership to the company and the industry. We could try to be more visible about what we’re doing and not just talk about successes but our failures too, because they’re often very interesting.”
Whether they’re talking about success or failure, the data science team at Brandwatch is a fascinating group of people with a broad range of talents and a dedication to solving often very complicated problems.
I might be interning on the marketing team, but data science are definitely the coolest team at Brandwatch.