What is the Certificate in Data Science, and how is it boosting actuaries’ data literacy? Manuel León-Urrutia explains
Data is the raw material of the actuarial profession – and it is increasing exponentially in terms of quantity, variety and ease of access. Often powered by artificial intelligence (AI), the tools to process data are fast evolving, too. Access to these tools, as well as the data that fuels them, is becoming more democratic, and through this democratisation, products and services gain value. This benefits all stakeholders, from providers to end users, which ultimately helps everyone in society. However, tech giants can hoard data and monopolise the tools that unlock its value, potentially leaving specialised professions out in the cold. In addition, mistrust of these innovative tools can prevent such professions from jumping on the innovation bandwagon. This mistrust often manifests in two main forms: the perception that these new tools can only generate byproducts, and the fear that they will replace specialists, leaving them out of a job.
Another challenge is the data skills gap. Recent reports such as Accenture’s Closing the Data Value Gap and the Data Skills Project’s The Human Impact of Data Literacy, reveal that a shortage in workforce data literacy is one of the main reasons that the value of all this data is not sufficiently realised. The actuarial profession is not an exception, even considering qualified actuaries’ rare and exceptional qualities.
Data literacy is the ability to critically understand data and extract value from it. This is partly inherent to actuaries, whose task is to calculate risks and make decisions based on data; actuarial qualifications ensure that members of the profession have the advanced numeracy required to carry out this task. However, data literacy is more than just the ability to perform sophisticated statistical analysis. The ability to generate narratives from data for different audiences, manage ethical implications of data use, and select appropriate machine learning techniques depending on the data available, are just a few of the many aspects of data literacy that should not be taken for granted in any data-driven industry.
In 2020, the IFoA and the Southampton Data Science Academy partnered to release the Certificate in Data Science, an online learning programme that is aimed at actuaries in all stages of their careers. The programme has been designed in response to the need to update actuaries on the latest developments in data science, big data, machine learning and AI. With nearly 400 actuaries in three cohorts trained at the
time of writing, the course has proven its success in enhancing participants’ competence in handling the complexities of such a data-driven profession.
Who is it for?
This programme is available to the whole range of IFoA members, from recent graduates aiming to pursue an actuarial career to experienced actuaries who have long completed their certification cycle
and want to engage in further professional development. Participants come from all industries in which actuaries have a role, such as banking, health and risk management – although there is a certain predominance in the insurance and reinsurance sectors.
A fairly frequent profile is that of an insurance professional aiming to pursue an actuarial career, taking this programme during a break in their official qualification route. Another frequent profile is the senior actuary aiming to become acquainted with the latest data science and AI technologies. The programme caters for all these profiles, providing individualised learning experiences and leveraging the diversity
of its participants.
What does it cover?
The programme covers data science and AI core concepts and common techniques in combination with actuarial case studies. It is divided into three main blocks. It starts by addressing the main stages of the data science pipeline, namely data management and preprocessing, data analysis, and data visualisation and storytelling. Secondly, it moves into specific AI methods and considerations. Thirdly, it explores the application of the two first blocks within the actuarial profession, with an emphasis on data science best practices. Specific actuarial case studies about the use of data science and AI are introduced throughout the course in relevant topic sections. The course also contains optional guided coding exercises in Python, addressing all stages of the data science pipeline.
“Data literacy is more than just the ability to perform sophisticated statistical analysis”
How is it designed?
The Certificate in Data Science programme is designed so that delegates attain a specific set of learning outcomes, namely understanding core concepts in data science and how they relate to AI and machine learning; retrieving, processing and managing relevant data in a range of formats; analysing relevant datasets with state-of-the-art tools and techniques; identifying opportunities to apply business solutions with the latest machine learning and AI technologies in an actuarial context; and gaining insights about the legal, ethical and technical implications of using big data and AI in an actuarial context.
The programme is cohort-based and fosters conversation between participants. This is its main feature: the learner has multiple opportunities to talk about topics with other participants and the tutoring team via discussion forums, group tutorials, chats, individual tutorials, assessments and messages. The course is a long conversation about data science that lasts 10 weeks. Even the assessment is a conversation, with the tutor making individualised comments about what the delegate produces, and the delegate having the opportunity to discuss the tutor’s feedback.
Who is behind the scenes?
A multi-disciplinary team of actuaries from the IFoA, data science academics from the University of Southampton and online learning designers from Cambridge Education Group Digital have worked together to create this learning programme. The team meets iteratively to address a feedback loop that fine-tunes the course before each new run, which launches three times per year.
How is it doing?
As of January 2021, a third cohort is being trained. The two previous cohorts have been mutually evaluated by both delegates and the education team, with generally positive results in both accounts.
Composed of higher education academics at the frontline of programme delivery, the tutoring team has been impressed by the levels of commitment, engagement and diligence shown by the programme’s participants. Before engaging in the course, tutors were aware that they were going to provide training to a profession imbued with solid professional standards, but they all agreed that their expectations of the participants was exceeded.
The feedback from the learning community has also been positive. Delegates highlight the programme’s ability to demystify data science and AI, unveil its main concepts and technologies, and enable participants to hold an intelligent conversation about the latest technological developments. Perhaps some participants expected the programme to emphasise coding more than it does, especially as the coding exercises are not assessed. However, this is generally compensated for by the participants’ ability to self-learn and the support provided by the tutoring team.
It is still too early to determine the programme’s impact on its participants’ professional development, but there are already indicators that the lessons from the programme are being successfully applied in some delegates’ day-to-day work.
For more information about the Data Science certificate please visit bit.ly/3vpzCUl
Dr Manuel León-Urrutia is senior teaching fellow at the University of Southampton’s Electronics and Computer Science department, and head of learning at Southampton Data Science Academy