As part of the launch of the Data Skills Taskforce's latest report, Data Skills for the Future, which is led by Accenture and Pearson, innovation centre Nesta convened a panel of data professionals to discuss data skills in education and the workplace.
The panellists were Dr Nicolas Guernion, director of partnerships at the Alan Turing Institute, Dr Richard T. Freeman, lead data engineer and architect at Just Giving, Dr Michel Wermelinger, senior lecturer of the faculty of Science, Technology, Engineering and Maths at the Open University, and Dr Kim Nilsson, CEO, Pivigo. The panel was moderated by Martin Squires, global lead for customer intelligence and data at Walgreens Boots Alliance.
Squires: Are companies looking for a mythical data scientist, someone with a raft of skills and decades of experience, or are they looking to fill other roles that have been placed under the umbrella term data scientist?
The consensus on the panel was that no data scientist is an island. Companies that want to be data-driven are demanding too much of prospective candidates, but actually need people with data skills to fill various roles.
Freeman said that, while the data scientist is the popular role that everybody talks about, the data scientist doesn’t work alone. He pointed out that data engineers and machine learning engineers automate the processes that feed the data that the data scientist consumes to create models. Also, business intelligence teams have a critical role in shaping the data and maintaining the data warehouse.
Furthermore, he said that it is important for people at management level to have an understanding of data skills and thus make the company more data-driven.
"The role of data scientist can be broken down into eight different jobs."
Nilsson commented that companies can be extremely demanding when it comes to the pre-requisites of candidates and, as a result, many say they can’t find data scientists, while many data scientists say that they can’t find jobs.
“One of the key challenges is that the industry is looking for the unicorn data scientist,” she said. Nilsson referred to the Big Data Skills Gap report by techUK which broke down the role of data scientist into eight different job roles, including the data analyst and the data visualisation expert.
She said that finding talent is about identifying exactly what type and at what level is needed and how they can combine different people, different skills and different backgrounds together.
Wermelinger said that the Open University needed data wranglers with good communication skills to process student surveys because they need to be able to explain the data and the trends. He also mentioned that the MK:Smart initiative, a smart cities project for Milton Keynes led by the Open University, needed a multidisciplinary team because it has an infrastructure data hub.
“You need people who worry about the hardware, develop the data privacy issues, legal issues and the business model after the project ends,” he said.
Squires: Is it necessary to have a PhD or a STEM background to work in data science?
The consensus view was that it is not necessary - passion is far more important.
Nilsson said categorically: “You absolutely don’t need a PhD,” pointing out that managers and data visualisers do not need to have doctorates. However, she did say that certain skills that one learns during a PhD are very advantageous when going into a data science role.
Guernion noted that, although it might be hard for someone with a social science background to migrate to a programming language and go into it with a lot of depth, that person would have other skills that could really benefit the “industrial context”.
"You need to be aware the data doesn't tell you everything."
Wermelinger made the point that several students who are enrolled on his Mooc (massive open online course) come from a wide range of backgrounds - one of his top students is a former journalist with a BA. According to Wermelinger, those who want to learn data skills need to be engaged by the data. They also need to be aware the data doesn’t tell you everything and be conscious of the assumptions behind data collection. In addition, communication skills are so important that essay writing has been added to the list of assignments for the Mooc.
It doesn't matter to Freeman if someone is not from a STEM background, as long as they are willing to learn the analytical thinking, the maths, as well as statistics and coding.
Squires: How can data skills be expanded? How can we increase the number of people with data skills in the workforce in the short, medium and long term?
By consensus, there is a need to provide resources to facilitate autodidactic learning, to highlight the interesting, exciting and personal aspects of data which would make a career in data science more appealing to young people, and to embed skills related to data science in the school curriculum.
Guernion said it would be really useful to have conversion courses so that someone with a background in a different subject, for example physics, could deepen their understanding of techniques like machine learning. He also said that apprenticeships could be used more flexibly and it is good that there is greater recognition that they can be used at even post-graduate level.
Freeman is a big advocate of autodidactic learning. He feels that, once an individual has identified what area of data science they are interested in and they have a passion for the subject, there is a wealth of resources for self-teaching, such as books, online courses, open source libraries, online competitions, and open data sources like the data.gov portal that they can learn from.
"Data skills are the future of the UK economy."
Nilsson said it is important to start early to educate young about data skills because, although universities are turning out more people with an MSc or PhD in data science, “it’s a drop in the ocean”. “We have to start very simple and get excited about data and about digital skills. This is the future of the UK economy,” she said.
Wermelinger as an educator has a special perspective on this issue and gave the fullest answer. He referred a project at Urban Data School, which grew out of MK:Smart, about solar panels and heat loss in Milton Keynes. With the information being close to home for the students, they could personalise it, pointing out familiar roads and comparing the data about their streets with those of their friends.
“Even starting with very young ages, engaging with data is very do-able,” he said. Wermelinger is pleased that extra-curricular activities around data skills are taking place at schools, but he feels that they need to be properly incorporated. He said: “Unless the curriculum in schools really changes to make space for data activities, teachers are going to be pragmatic and only teach the curriculum."
He has heard from his colleagues at the Open University that the new maths curriculum will not make much difference at GSCE level because there is not much maths handling, but at A-level, students really have to learn a data set because they will be assessed on it.
“Unless there is a holistic view of data in the curriculum, it is not going to be transformational.”
He also said it will be important to educate the educators because statistics is now mandatory at A-level, so the Open University has modules to enable maths teachers to pick up key statistics skills.
Wermelinger concluded with an example of how a data skills project could be taught across multiple classes. He said that in PE, one half of the class could be counting the passes of the other half that is playing football. In maths class, they could learn how to analyse the data and in art class they could learn how to visualise it. And then they could return to the PE where they could interpret the results which requires subject knowledge. He concluded: “Unless there is a holistic view of data in the curriculum, it is not going to be transformational.”
Caroline Florence of Narrative Insights is speaking on "Busting the skills gap myth and taking control of the problem" at DataIQ Future on 19th October. Book your ticket here.
Thank you for your input
Thank you for your feedback
DataIQ is a trading name of IQ Data Group Limited
10 York Road, London, SE1 7ND
Phone: +44 020 3821 5665
Registered in England: 9900834
Copyright © IQ Data Group Limited 2024