While some industries are early adopters of data science into their processes, all industries will eventually use data sciences as a core part of efficiency and optimisation tools, says data science training organisation Explore Data Science Academy (EDSA) commercial officer Mark Schroeder.

Banks and retail and telecommunications companies are early adopters because they hold significant amounts of data that are typically only lightly used, he says.

“The mark of modern business intelligence is not how much you know, but what questions you can ask. “This requires not only data, but also data science and domain knowledge, inherently making these projects collaborative in nature,” he avers.

The nature of data science is to develop quantitative processes to solve problems or to derive more insights to inform business processes and personnel, says EDSA head of faculty Dr Jaco Jansen van Rensburg.

“Data scientists come from a range of disciplines, including the engineering, actuarial, statistical sciences and mathematical disciplines,” he says. “The academy was founded by data scientists who have actuarial, consulting and academic backgrounds,” he adds.

The academy has more than 350 students this year and is well on track to meeting its target of 500 students by next year. This is the result of expanding its footprint outside Cape Town, with a new facility in Braamfontein, Johannesburg. EDSA is also considering expansion into other regions, he says.

The first phase of data science training that the academy focuses on is digital skills and familiarising students with the tools that they would typically use. This phase also includes training on cloud systems and common data environments of companies.

“Problem-solving capabilities are, obviously, a core component of the data scientist skill set, but training must include the approach to a problem and the method to be followed, as well as thinking about the code and structure of any solution to be able to effectively address the problem,” explains Jansen van Rensburg.

The second phase of training involves the students collaborating to solve real problems presented by sponsor companies. This ensures that students can immediately start solving problems when employed, but also highlights domain-specific complexities that students must navigate.

“Part of this training is instilling a sense of enablement where students are given various resources and leads and have to teach themselves how to solve specific challenges. “This not only prepares them for the variety of problems they will grapple with once they are working but also enables them to steer their own career paths in the directions they find engaging or compelling.”

Schroeder highlights that the academy is solving real-world problems, excluding those served by its data science consultancy subsidiary, Explore-AI.

“The problem must be recognised at the highest echelons of an organisation’s leadership and then a return-on-investment case must be developed. This will determine whether the problem justifies a data sciences approach.”

Thereafter, a domain specialist, usually from within the company, must serve as a data translator – a person who can explain the relevance of the organisation’s processes, which, in turn, enables the data scientist to determine the weighting that data sources carry in the models.

Companies can prepare to leverage their data more effectively by ensuring that it is as clean and organised as possible and, more importantly, that it is accessible, typically through application programming interface so that the algorithms can be applied to it, says Schroeder.

The skilled graduates the academy produces often form a key part of digital transformation initiatives in organisations, and are geared for working in the digital economy, he concludes.