Hirer beware: The data scientist pool is swimming with pretenders

In the past week alone, the ABC, South32, Avanade and Woolworths have each posted job listings for data scientists.

The vacancies are likely to attract a substantial number of applicants. Last month, job site Indeed saw a 138 per cent surge in searches for data scientist jobs in Australia, a trend echoed worldwide. The role saw the highest spike in local searches after only 'nurse' and 'teacher'.

But many of those sending in their CVs for what has been dubbed “The Sexiest Job of the 21st Century” will not have what it takes. Worse still, they might be frauds.

As one industry insider put it, the data scientist talent pool is swimming with “pretenders”.

Awesome bandwagon

More and more Australian organisations are building in-house data analytics capabilities, as they look to capitalise on big data.

At the Gartner Data and Analytics Summit in Sydney last week, Woodside data science manager Shelley Kalms described how its data team had expanded to 20 staff since it was set up in 2015.

Darren Abbruzzes, general manager of data at ANZ – which appointed its first chief data officer earlier this month – shared how employees were being given the tools to do their own analytics. NSW minister Victor Dominello explained that the state government’s Data Analytics Centre had grown from “one man and his pot plant” to a 35 staff strong “powerful beast”.

“We are past the days when finance, marketing, and defense were the only options for these applications,” explains Dr Hercules Konstantopoulos, who has recently joined Atlassian as a data scientist from a similar role at energy software solutions company Envizi.

“Software companies are now big attractors for people who will build models of their product usage or growth; non-profit organisations are using advanced analytics to find the right supporters or volunteers; and the public service is advertising positions in an array of areas that help shape governance and policy. With every month that goes by a new area of business enters the realm. Every industry is investing in this awesome bandwagon.”

Konstantopoulos, an instructor at General Assembly and former astrophysicists, adds that to those businesses with big data, data scientists are considered to be “the custodians and gatekeepers of this realm”.

But buyer beware. Ben Davis, senior architect at Teradata serving the federal government market in Canberra, argues in a recent blog post that the term data scientist is “over-used and even abused”.

The industry is being “flooded with fakes all looking to get in on the action”, he wrote, describing the situation as akin to “the wild west”.

Unicorn smiles

A recent Institute of Analytics Professional Australia (IAPA) survey found that business managers are searching for “unicorns”: Candidates with the necessary analytics and statistics skills and coding competence, matched with an ability to communicate with various business functions and drive outcomes.

“Data scientist roles are usually well subscribed but many candidates don’t quite have the combination of technical skill and business competence,” head of data excellence at AGL Energy, Buzby Kuramoto, told the IAPA.

Data scientists need to be proficient in data exploration and query languages; data analysis and scripting as well as programming and collaborative coding. They need the ability to design new algorithms, handle big data, and have some domain expertise. More than this they need to interpret and explain their findings to different functions of the business.

“Typical tasks involve exploratory data analysis, building statistical models, and putting together compelling data visualisations and presentations. Data scientists tend to be called in to interpret and explain data, trends, and information in abstract and open-ended ways, which creates a bit of a distinction from more traditional roles, such as data analysts or business analysts,” explains Konstantopoulos.

“The job of a good data scientist is not only to solve problems, but also to discover the questions worth asking.”

Depending on the size of the business, data scientists may report to a chief digital officer (CDO), sit in a consultancy-style, business-wide analytics team, or, as is often the case, report to a technology function or CIO.

“Which is not always ideal,” Konstantopoulos adds. “Most data scientists report to non-specialist managers who think they are made of rune magic and unicorn smiles.”

Not all are so easily dazzled. The IAPA survey found that almost a fifth (19 per cent) of business leaders questioned said they were “slightly unhappy” with their resident analytics professionals.

Noise in the pool

The attraction for chancers is understandable. Data scientists have the opportunity to work at the forefront of emerging technologies like AI and machine learning and the chance to make a difference in the world. They also command high salaries — averaging $130,000 (according to the latest IAPA salary survey).

‘Weeding out the fakes’, as Davis puts it, is hard. “It is often difficult for an employer to identify how good they say they really are,” he wrote. This sentiment was supported by the IAPA survey findings that found that employers were frequently unable to independently verify a candidate’s professed skills in the field.

“In terms of supply, it is quite difficult to find the right person for the job,” says Konstantopoulos. “There is no formal certification, no data science guild, so anyone can assume that title. As far as employers and recruiters are concerned this increases the noise in the talent pool rather than increasing availability, so there is a bit of a shortage of people whose skills live up to the title.”

The value of a good data scientist to a business is huge. There are numerous examples of how those working in the role have succeeded in efforts to unlock enormous benefits. However, as the demand for data scientists grows, so will the difficulty in hiring the best ones.

“It takes years of training to understand how to wrangle data,” says Konstantopoulos, “and that specialisation is not likely to go away as businesses collect more and more information.”

As Davis puts it: Finding the best is “harder than unearthing a pink diamond in your backyard”.