Riaan de Jongh tips for bridging the skills gap
The data science skills shortage has expanded. This is mostly because of the increased demand for analytical skills, rather than there being fewer data scientists. But this is still a problem.
Traditional methods—particularly training more data scientists via university courses—have proved insufficient to meet demand. Companies are turning to alternative means such as democratization of analytics, improving employee skills and developing partnerships with universities. But are there aspects of the situation that we are missing?
Riaan de Jongh is a retired professor, now on contract at the Centre for Business, Mathematics and Informatics (BMI) based in Potchefstroom, South Africa. He is a strong believer in the power of innovation, idea generation and creative problem solving, and recently joined us for a #SASChat to discuss the skills gap in data science. Some fascinating insights and ideas emerged.
A growing demand
We started by talking about why demand was still growing for data science skills. Riaan suggested that the demand is driven by many factors. He cited the immense increase in computer processing power and the growth of the Internet of Things (IoT), the availability and expansion of more and more data sources, and the rapid developments in artificial intelligence and machine learning. He went on to say that these developments are creating new business opportunities, but they can only be exploited with the help of skilled data scientists.
The #SASChat noted that there are some real risks to both government and business if this skills shortage is not addressed. Riaan commented that skills shortages will limit the ability of companies to exploit many of the existing and new business opportunities, which could result in a loss of their competitive edge. The problems that are likely to emerge could be across various areas, including the range of services or products, the effectiveness and efficiency of operations, and potential cost savings and profit generation.
The real question, therefore, is how we can reduce the skills shortage. Riaan suggests that the answer is a different approach to teaching, saying: “We need to change the mind set of data science students to become professional problem solvers and lifelong learners.”
He noted that at the Centre for BMI, students are taught these skills. They are given real-world business problems that stretch them, and are taken through a process of problem formulation, solution generation, project planning and execution. This helps them to develop resilience and creative thinking. These skills are essential, because data science is all about asking the right questions. He cited William Edwards Deming’s comment that “If you do not know how to ask the right question, you discover nothing.”
Riaan also quoted Bradley Efron, President of the American Statistical Association, as saying,
“Scientists have misled themselves into thinking that if you collect enormous amounts of data you are bound to get the right answer. You are not bound to get the right answer unless you are enormously smart.”
Driving change
How can businesses and government drive change? Riaan suggests that primarily, data scientists need to be motivated to develop their own skills through involvement in real-world projects. However, there is also a place for partnerships between industry and academic institutions, for example, through the provision of funding for data science courses, or bursaries for students. This would allow businesses more say in the teaching, and therefore give academic institutions a better understanding of what industry requires from its data scientists.
Riaan commented that the best upskilling technique he has seen is work-integrated learning, where academics work with practitioners on projects to solve business problems. This must be under the leadership of an experienced business professional who owns the problem. However, there are issues with this kind of partnership. For example, academia and business have very different reward structures and performance measures, and these can sometimes come into conflict. They also have different objectives, and this can lead to misunderstandings and issues of prioritization.
The tweetchat also identified other solutions that may help to close the skills gap. Federica Citterio noted that tools can help the democratization of analytics, for example by providing low-code or no-code interfaces. Commercial software can therefore play an important role in this process. However, perhaps the last word should go to Albert Einstein, via Riaan:
“Information is not knowledge, the only source of knowledge is experience.”
Every Friday, there's a new edition of our #SASchat on #twitter - join us.
Just this Friday, a great panel is talking about: "Seasonal schools on the path to democratization of AI" - how you can be part of it? - All info here: Seasonal schools on the path to democratization of AI | SAS Let's meet there!
#DataScience #AnalyticsLeadership #DataScienceSkills
