Registered users can save articles to their personal articles list. Login here or sign up here

Solving the data science conundrum

A new academy has a novel approach to training students.

South Africa is facing a shortage of data scientists – a new breed of analytical data experts with the technical skills to solve complex problems. They’re part mathematician, part computer scientist and part trend-spotter. And, because they straddle both the business and IT worlds, they’re highly sought-after and well paid.

The demand for data scientists is being driven by the emergence of big data – that unwieldy mass of unstructured information that can no longer be ignored and forgotten. It’s a potential gold mine for companies – as long as there’s someone who can dig in and unearth the business insights that no one thought to look for before.

South African universities like Wits and UCT have introduced Data Science degrees at the Masters level, but this is producing about 40 data scientists a year – far short of the number that the country’s banks, insurers, retailers, health companies and telecommunications providers, among others, require.

So three actuaries Shaun Dippnall, Aidan Helmbold and Dave Strugnell came together to solve the problem in the way that actuaries do – by finding an innovative solution.

“South Africa is facing an enormous unemployment problem at the same time as the world is heading at breakneck speed towards the Fourth Industrial Revolution,” says Dippnall.

The possibilities of billions of people connected by mobile devices, with unprecedented processing power, storage capacity, and access to knowledge, are unlimited, he says. These possibilities will be multiplied by emerging technology breakthroughs in fields such as artificial intelligence, robotics, the internet of things, autonomous vehicles, 3D printing, nanotechnology, biotechnology, materials science, energy storage, and quantum computing – even in South Africa.

“This means that more jobs will be mechanised at a time that we desperately need to be creating jobs or reskilling people.”

So together the founders created an academy, the Explore Data Science Academy, that will create data scientists at velocity; that has no qualifications for entry and that does not teach using the traditional classroom, ‘sit still and listen to me’ approach.

The academy is in its fourth month and most of the hundred students that joined in February are progressing well. Thus the process has begun to recruit four times the number of students for the 2019 intake and find the companies willing to sponsor them.

“The recruitment process is non-traditional in that we require no prior qualifications; instead we do a number of tests that measure natural raw aptitude – specifically the ability to program, solve problems and work with numbers,” Dippnall says. “We also look for the ‘light in their eyes.” This, he says, refers to a hunger to learn and curiosity about the world around them.

About 10 000 people between the ages of 18 and 35 applied to join the course, 250 were interviewed and 100 were selected. Of these, 25 have no prior learning – just matric with maths and science. “There are so many talented people out there – and these were not kids from former Model C schools. It proves the point that all our kids lack is opportunity.” 

On day one the students are thrown into the first of a series of ‘sprints’. In each sprint students are grouped into teams of five and are given a problem to solve – for example “predict someone’s personality profile using just their Facebook posts”. They have three weeks to build the algorithms that would complete the task.

A full-time team of 14 academics (including statisticians, engineers, actuaries and physicists) teach the mathematical, statistical and computer science fundamentals required. They also provide students with the latest machine learning techniques and data science tools (Python, SQL, Power Bi). “In this approach the problem is primary, while the theory is secondary,” he says.

The course is comprised of six months made up of the three-week sprints described, a three-month project and a three-month internship. In this, the founding year of the Academy, IT company BCX has provided R50 million in sponsorship and intends taking on all 100 candidates for their internships. It may also employ some once the course is complete. The sponsorship is spread over three years and covers the cost of study for 100 students each year, as well as a small stipend for financially-needy students.

For BCX the advantage of the programme is that it increases the flow of data science skills into the business and executives are exposed to a veritable pool of talent that have developed practical problem-solving skills that businesses need.

In addition, the course is an accredited skills programme which means that it can claim BBBEE points from the spend.



To comment, you must be registered and logged in.


Don't have an account?
Sign up for FREE

This is awesome.

Is it really awesome? This will be another failure for a profession and lead to a continuation of business as usual in SA.

Let me explain, specialist in every profession exists and require years of study but in my short life time ive seen businesses “train” people to do the job of a specialist at the expense of a proper specialist in the name of short cutting the production line and producing the much needed skills. This is often tied with social good initiatives for transformation or poverty elevation etc. but the brutal truth is this is reducing the cost of specialists, lowering the bar as it were while causing existing specialist to be overworked as managers for these new guys to lower the cost.

I’ve seen this happen with engineers, computer science and now specialist computer science? Fun times ahead.

Ps the problem with SA economy is we a developing economy but it only benefits managers.

I agree.

We know the education system in SA is the biggest part of the problem. but there is no short cut,

Wow!! and what a great mistake it is to get people to just apply ‘big data’ methods. These ‘big data’ methods are garbage in garbage out. I can think they will be good economists though.

But who knows how this business as a ‘Academy’ is conducted and financed, maybe even with some government backing also..

Sad but even the university education these days is such a funny business, government and academics pushing agendas giving them worthless pieces of papers.

Stellenbosch University recently published a list of curricula that will be decolonized…Even engineering is on the list.

Another point I have a problem with is hearing that students work as teams. That then results inevitably in a case of the team having natural leaders and some bright members who benefit the less capable in the team. Yet the whole team receives the same credit. Passengers need to be identified and graded as such. How can that be done in a team assignment?

R500k per student for the equivalent of a 3 year degree?

Not really as it is not a formal qualification..?

“Sit still and listen”?

Does not sound very innovative or cost effective.

That said, this country needs all the help it can get to educate people.

Would be interested to know if these actuaries are doing this in a charitable manner, or if not, what they stand to gain in terms of compensation..

I do like the “light in their eyes” bit though.. Need some hungry, driven individuals.. However, sometimes the lights are on but nobody is home.

Cool idea. But there are literally no actual qualified data scientists involved it seems. Sort of the blind leading the blind imo

NWU is leading the way in this field. There is simply no shortcuts to a proper degreed qualification in this field. Can any multimillion Rand company get away with a wrong decision if they say they self taught analysts made the decision? I don’t think so.

This sounds like those 6 week “MBA” qualifications.

just like the “GMP” courses all the big banks and insurers send their people on so they can put Harvard on their CVs. Worthless.

A clever way to get young students to do the tedious bits and bytes of big data. Analysing big data requires hundreds of hours of painstakingly building/populating databases and then deploying queries to extract specific results – its not rocket science you jus need someone to the spade work.

Load All 10 Comments
End of comments.


Insider GOLD
ONLY R63pm

Moneyweb's premium subscription is a membership service which will give you access to a number of tools to take charge of your investments.
Or choose a yearly subscription at R630pa - SAVE R126

Get instant access to all our tools and content. Monthly subscription can be suspended at any time.



Follow us:

Search Articles:Advanced Search
Click a Company: