This is the first iteration of an actuaries’ data science wiki. The aim is to expand, refine and develop this wiki to encompass a knowledge-base that scopes the key terms and essential definitions of disciplines associated with data science practice, that have particular resonance for actuarial professionals.
Data analytics Data analytics is the discipline of analysing data sets to make conclusions about that information. Data analytics techniques can reveal trends and metrics that would otherwise be undiscoverable in massed information. This information can then be used to optimise processes to increase the overall efficiency of business or system operations.
Data analytics is a broad term that encompasses diverse types of data analysis. Any type of information can be subjected to data analytics techniques to gain insight that can be used to achieve improvements. Some of the techniques and processes of data analytics have been automated into mechanical processes and algorithms that work through raw data for subsequent human analysis.
Data analytics methodologies include exploratory data analysis (aims to find patterns and relationships in data), and confirmatory data analysis (applies statistical techniques to determine whether hypotheses about a data set are true or false). EDA is comparable to ‘detective work’, while CDA is comparable to the ‘work of a judge or jury during a court trial’ (John W. Tukey, Exploratory Data Analysis, Pearson, 1977).
Data analytics can also be separated into quantitative data analysis and qualitative data analysis. The former involves analysis of numerical data with quantifiable variables that can be compared or measured statistically. The qualitative approach is more interpretive – it focuses on understanding the content of non-numerical data like text, images, audio and video, including common phrases, themes, and points of view.
Data analysis, data analysts Data analysts and actuaries share similarities. They have comparable skill sets, and use mathematics, statistical techniques, and computer knowledge to compile and analyse data, and to report conclusions for business decision-making. The two disciplines differ in the scope of their work and employment settings.
For instance, data analysts work in a broad variety of vertical sectors and industries with multiple types of data. They apply mathematical and statistical techniques to extract, analyse and summarise data. They use spreadsheet and statistical software, work with relational databases, and prepare charts and reports of their findings. Their work transforms large, complicated data sets into usable insights that inform organisational leadership decisions and policies.
Data analysts review information and use the data to help develop recommendations. They do not specifically focus on risks, but may help determine appropriate business or financial decisions that will benefit a company.
Data visualisation The main goal of data visualisation is to communicate information clearly and effectively through graphical means. and by maintaining a library of data visualisation techniques. The IFoA Data Visualisation Working Party was established in 2017. Its vision is that data visualisation for actuaries should represent:
- An understanding of which visualisations work well for different purposes.
- Domain-specific examples of helpful practice.
- An understanding of how to produce the visualisations, including tools and techniques.
- An understanding of the principles of developing and improving data visualisations.
- Awareness of caveats that should be associated with data visualisations.
Machine Learning Machine Learning is a discipline that uses study of algorithms and statistical models, as used by computer systems, to perform specific tasks without use of explicit instructions: Machine Learning instead relies on patterns and inference. It is generally regarded as a subset of Artificial Intelligence.
The question of what Machine Learning could bring to actuarial work is something of a contentious issue within the insurance sector. Some have speculated on Machine Learning’s capacity to replace manual actuarial work, and therefore reduce insurers’ requirement for human actuaries. Other argue that data science-savvy actuaries could turn knowledge of Machine Learning into a useful asset in their skills offering.
Predictive modelling Predictive modelling involves the use of data to forecast events. It relies on the capture of relationships between explanatory variables and the predicted variables from past occurrences, and the exploitation of this to predict future outcomes. The forecasting of future financial events is a core actuarial skill. Actuaries routinely apply predictive-modelling techniques in insurance and other risk-management applications.
Filter or search events
Louise Pryor, IFoA President, will chair this free-to-view session, in which Alex Darsley, TPR, Actuarial Regulatory Policy, will be discussing the regulator’s Climate Change Guidance Consultation, which is seeking views on new guidance designed to help trustees meet tougher standards of governance in relation to climate change ri
This session will examine the megatrends and themes driving the Future of Work across the Financial Services industry, and how Covid-19 has accelerated new future work priorities with a particular focus on hybrid work and leadership mindset and capabilities.
Internal audit is often the Cinderella of the audit world. It’s a regulatory requirement for insurance companies to have an internal audit function, so why not make it as useful as possible? This session will look at how to link an internal audit plan to the risk register, and how that helps audit committees and boards to spot problems and fix them.
Climate change is one of the greatest risks facing our world today. Addressing it will require multi-faceted solutions. Through this panel session, we will explore the different levers that can be used to meet net-zero targets including climate science and data, government engagement, and mobilising green finance.