In 2009, Google’s chief economist, Hal Varian, trumpeted the future value of people with statistical skills. He said: "The sexy job in the next 10 years will be statisticians… The ability to take data, to be able to understand it, to process it, to extract value from it, to visualise it, to communicate it.” Seven years on, there is no denying the demand for these skills. However, there is debate over what to call the people who possess them. Are they statisticians, data analysts or scientists?
Last year, David Donoho, a professor of statistics at Stanford University, drew attention to what he called “today’s data science moment”, the phenomenon of which “the statistics profession is caught at a confusing moment”. Donoho believes the work conducted by statisticians for centuries has been thrust into the limelight but these activities are being viewed as new and carried out by “upstarts and strangers”.
What it boils down to, according to Donoho, is the similar definitions of what a statistician does versus today’s data scientist, a comparison that way back in 1997 led Professor C. F. Jeff Wu, the then-Chair in statistics at the University of Michigan, to call for statistics to be renamed data science and statisticians data scientists.
The American Data Science Association defines a data scientist as “a professional who uses scientific methods to liberate and create meaning from raw data” while statistics is “the practice or science of collecting and analysing numerical data in large quantities.” The difference between the two, according to Donoho, is that the definition of a stats seems “limiting”. From this, he draws the conclusion statistics is being marginalised in the race to champion data science.
Louise Ryan, Professor of Statistics at the University of Technology Sydney, says: “When you read about data science or you look at announcements for conferences and so on, you often don't even see the word ‘statistics’ appearing. It's data science, analytics or machine learning. Sometimes in the statistics field, we're like, what about us? Aren't we important?”
Statistical skills are clearly important but those in the field say they need to be paired with additional knowledge. Kendra Vant, General Manager of Big Data Analytics, Big Data Global at Telstra, says: “If you look back at where the name data scientist came from – and it hasn't actually been around that long – it is someone who knows more statistics than a computer programmer or more computer programming than a statistician.”
Vant says the key differentiator for today’s data scientist is the ability to write in scripting languages such as Ruby or Python, or more niche languages along the lines of R and SAS. Still, Vant says locally we are lagging behind countries such as the US when it comes to producing professionals with skill sets of this nature. She says: “I don't think we see that quite as uniformly across the Australian market yet.”
UTS’ Ryan agrees that a broader knowledge base is required to work in the field. She says: “If you just study statistics, that's probably not quite enough. You need a modern blend where you've got training in statistics and some training in computer science because these days it's not just the stats methods. You have to be really good on the computational side. You have to know how to work with big databases. You have to know how to work with internet data and distributed data systems. You have to be able to communicate. We really need a transformation of statistical education so that kids come out with the right blend of skills to be able to move in all these areas. If you get somebody who's got those skills, they can be helpful in virtually any industry.”
While the difference between a statistician and a data scientist is still being ironed out, Vant says there’s also another shift in role designations. She says: “There are a number of people who call themselves a data scientist now that two years ago would have called themselves a data analyst.”
Strength in numbers
Semantics aside, the industry is clearly experiencing a surge. Vant says: “There's no doubt a lot of people want to get into this line of work. There's an absolute fascination with the field.”
Regardless of how it is labelled, having a statistics, data science and analysis skill set today opens up a world of career opportunities according to Antony Ugoni, Global Director of Analytics and Artificial Intelligence, at online employment marketplace SEEK. He says: “If I reflect on my career, I've worked in medical research. I've worked in fraud prevention at National Australia Bank. I've worked in marketing. I now work in the human capital market. I have no education or background in any of those industries, yet I've been able to have successful careers in each of them because analytics has become such an enabler of hidden value in those environments that you can pick up and drop into each of these industries and be really effective by applying rigorous analytic discipline. One hundred years ago, it’s probably oversimplifying it, but people like me, our career choices were to become a teacher or an accountant.”
Alex Concannon, national marketing analytics director for media agency Maxus, wasn’t even aware of media agencies before he started working in the field. He says: “I was a research scientist. My background is in bacterial genetics. I needed a job and ended up working in a creative agency and found that there was a niche I could carve out because I have an understanding of statistics, maths, and Excel. Then I worked for a couple of startups and tech businesses that were very data-driven. From there, I ended up at an agency that was looking for someone who could bridge the gap between creative and data.”
While media seems a long way from bacterial genetics, Concannon says most people with his experience tend to end up in financial services organisations but the culture within businesses of that nature isn’t for everyone. “I love working in media and the reason I've stuck around is it has a work hard, play hard culture,” he says. “I couldn't think of anything worse than working in a quiet, grey office, wearing a suit in some horrible corporate environment.”
Developing data literacy
As more and more businesses look to put data to work, it’s not only the data professionals that must skill up, management and leadership teams within organisations need to develop knowledge in the area in order to get the most out of the people working for them.
Telstra’s Vant says: “I believe we need to increase the analytical literacy of middle management and senior management across Australia. At the moment, Australian executives take trips to Silicon Valley, to Israel, to some parts of Europe which have fantastic data cultures and startup environments and they'll come back and say, ‘I've learned all this new stuff about how analytics can make my business smarter and how analytical companies make more revenue than non-analytically driven competitors. Now I want this in my company.’ Which is fantastic. But as analytical professionals, we then need to support them to understand both the upside potential and the required investment in treating data as an asset, data pipelines, data quality, etcetera.”
Vant believes anything that can raise awareness among mid and senior management about the power of analytics is going to help to transform the industry as leaders come to terms with the fact that going down this path is a long term strategy requiring the same level of investment as a new product line or manufacturing plant.
Maxus’ Concannon concurs that education is needed for those working with and managing analytics professionals, particularly in regards to the results they can deliver for a business. He says: “There's a bit of a misconception that data and analytics is a black box that just gives you answers. That's something which we've been working to educate people on. It doesn't give you an answer. It takes a large amount of information, processes it and then it digests it into a format you can use. At the end of the day, it still requires someone on the other end to interpret it and work out what that means for the business.”
Still, ambiguity might be a good thing for those in the field. Concannon says: “There has been a bit of an incentive to almost keep it mysterious because it protects you and inflates your value. There needs to be more sharing and explanation of what we're doing, why we're doing and how to use it so that we can educate the rest of the business as opposed to keeping it as a siloed business unit.”
As those in the field work to define their roles, educate and demystify the work they are doing, the question remains as to whether the profession can be labelled ‘sexy’. Concannon from Maxus says it can citing “geek chic”, while SEEK’s Ugoni says: “People are wanting to kind of hold up their data scientist badge loud and proud. It's a great time to be a nerd.”