According to techUK, the shortage of data skills in the UK is costing the country £2 billion a year. As a result, the industry is looking to bring in qualified candidates wherever they might be. Data science hub Pivigo is helping to plug this skills gap by training scientists to become data scientists.
The first five-week programme took place in 2014 and has since expanded to include an online course. The courses give those with doctorates and master’s degrees in analytical science the skills to work in industry. However, there are some considerable differences between working, researching and studying in academia, and working in industry towards a commercial end.
For Laila Alabidi, a data scientist at a startup and a former research fellow with a PhD in theoretical cosmology, the hardest aspect of the transition to industry was getting used to the culture of collaboration. She was accustomed to being self-sufficient and working independently and autonomously.
In academia, with everyone being focused on their own research, she would rarely, if ever, raise her hand to ask for help. “I always had the mentality that, if you ask for help, then you show your weakness, then maybe you don’t get a job afterwards,” she said. However, in stark contrast, when working in industry - and especially in startups - she found the opposite to be true where the culture has been “really awesome”. She elaborated by saying that requesting help is actually encouraged as colleagues are happy to point each other in the right direction.
She said: “In industry, you have to give up on the concept of being the star. At the same time, you gain the benefit of not having so much pressure on your shoulders. The ‘publish or perish’ mentality, in my case, killed creativity instead of nurturing it and I found in industry I was able to play around a bit more.”
For Jonathan Brooks-Bartlett, a data scientist at a large corporation and former post-doctoral researcher with a DPhil in mathematical biology, his scientific training was somewhat counter to the approach taken by data scientists in general. He said: “Falsifiability of hypotheses is not always a common concept that is pursued in the data science community as a whole.” Falsifiability is the principle that one conducts an experiment with the aim of disproving rather than substantiating it. However, data scientists will often test whether a hypothesis is likely to be true.
As well as culture, momentum is another perceived difference between academia and industry. Mike Bugembe, chief analytics officer at social giving platform Just Giving, said that, in his experience, new hires with PhDs need some time to adjust to the new tempo of working in industry and become commercially aware.
He said: “Academia works at a different pace to the commercial world.” In the commercial world, he commented, there isn’t the time or freedom to go off on a tangent and investigate something just because it is interesting or it increases knowledge without having a commercial application.
“It’s the Pasteur’s quadrant debate, the tension between doing something for increased knowledge and understanding, which academia can tolerate, versus doing something for commercial use and value.” He remembered working with one new hire with a PhD who was “fantastic” and “brilliant.” It took that person about six weeks to understand how different the commercial world is, but once they got to grips with the different culture, they were flying.