Put simply, data science is the application of scientific methods to business data to effect action. Data science includes technological tools and statistical methods to provide information to management to assist them in decision making and risk management. It enables us to extract insight from data that can lead to innovation and competitive advantages. It can also be a useful tool to develop an early warning system that picks up problems before they become too big to handle.


Who are Data Scientists

Data Scientists have a deep understanding of the data they work with as well as the tools available to them to analyse the data and draw insights from the data. They are able to build predictive models and analyse the results. They are curious, logical and have a solid base knowledge of linear regression and the associated statistical concepts. Many physicists, engineers, mathematicians, programmers and statisticians have embraced the tools available in this field. The accessibility to online material and instant access to learning material allows almost anyone to get started.


What skills do Data Scientists need

A good data scientist will not only be comfortable with programming and maths but have excellent communication skills to convey the implications of the use and results derived from the data. Communication skills are needed to understand and discuss clients’ underlying challenges and convey the results and implications of a model rather than just focusing on the details underlying the algorithms.

Data Science as a profession?

At this stage, there is no formal professional body that oversees the quality of education and output that all Data Scientists receive and produce. Thus there is a danger that techniques are inappropriately used, uncertainty is underestimated, and the weaknesses of the models are not quantified. Data Science work needs to be done properly, and Data Scientists need to make sure that data sources are clear, that transformations are tracked and evident, that data values are credible and their meaning is understood.

Communication skills are needed to understand clients’ challenges and convey the implications of the data used and the models used, at the right level to the right audience.


Data cleansing and exploration are crucial aspects of the process that are often glossed over.

As actuaries, we can take data science to the next level. With our professional training and skillset, combined with the toolkit that data science brings and the power of accessing and interpreting Big Data we can continue to add significant value to organisations.

Opportunities for Actuaries


In a way, a lot of the techniques actuaries use sit within the field of data science. After all, insurance is built on data. Actuaries, together with Data Scientists can provide value through research and the development of new models that incorporate new types of data.

There are so many fantastic opportunities for actuaries to get involved with, in the world of data science.

Actuaries can look beyond the data and make sure the numbers make sense and the risks and implications of the results are interpreted and clearly communicated. They can assist firms with the industrialisation of processes to make them automatic, secure, scalable and reliable. Actuaries also provide professional skills around communication and business conduct. Continuous Professional Development requires actuaries to keep their knowledge up to date.

Efficient Data Science is tailored to the data and the problem, and actuaries can assist with understanding the specific issues. Actuaries can be used to test and analyse models before they are put into production. Data Science is creating new toolkits for actuarial problems, and so programming is becoming essential for the actuaries of tomorrow.To get started in Data Science, an actuary needs to brush up on programming languages such as R, Python and a modelling platform. With so many resources available online an actuary can get started right away. One bit of advice is to start with a project in mind and learn by doing. It is also recommended to learn how to tell a story with your results and not just leave them as charts and graphs. And ensuring that, as actuaries, we continue to apply our professional conduct in every piece of work we get involved in.

All traditional actuarial techniques are already Data Science techniques. For example General Linear Regression Time Series and Bayesian Credibility Models.

Big Data?

Companies have been set up to store data from various sources and re-package it so that it can be analysed. Today, however more and more data are being generated on a constant basis. Computing power is making it easier to access and analyse these datasets at greater speed. Some of the benefits of working with richer data could lead to extra statistical significance and more potential correlations that can be explored. New data can be used to create new insurance products and open new markets.

The main risk is that in the pursuit of larger datasets, and without the underlying domain knowledge of the data involved, companies sometimes combine irrelevant data together which can distort the results of any analysis. For example combining data on car insurance where the one source is luxury vehicles in Europe and the other is second hand freight trucks in Africa, and then using the information to price car insurance to your average American family sedan.

Within organisations there are many sources of data, direct sources such as number of sales per month is already being processed but indirect sources such as how long each user spends on the company’s website can also be used to determine demand.Companies that invest in utilising and analysing better data sources will generate the most value and insight from this already valuable asset, which is their data.There are also external indirect sources of data, one example being social media, the use of which in insurance has been debated significantly. The New York Department of Financial Services is going to allow life insurers to utilise users’ online data when setting premium rates.. Some companies already turn to this data before hiring candidates.



Actuarial Case Studies

Insurance products are built on data. With the rise in technology, more data is being generated and accessed on a daily basis. Data Science not only helps insurers handle all this new data, but it also provides new tools to extract new insights. These insights can provide more information about customer churn, attrition and retention. This can supplement existing models to better price products and know which target markets to focus on. Actuaries can also start incorporating new datasets in risk models that may help them make more accurate predictions.

Looking to the Future

Increases in computer power are already allowing Data Scientists to automate repetitive manual tasks such as finding and preparing data. This automation will extend to more complex functions which will, in turn, save time and make more efficient use of firms’ resources. Software and hardware will continue to become cheaper and more powerful and more and more sources will be connected through Application Programming Interfaces. This will allow for the generation of Big Data sources where countless new insights can be discovered. Companies will compete to curate and ensure only the highest quality of data is allowed in.

More powerful platforms will host, maintain and secure these datasets in the Cloud. The Cloud will also allow more firms access to the latest hardware and software to crunch through this data via the Internet. Data Scientists will be able to orchestrate massive data analysis and see the outputs on their mobile phones. As new technologies emerge, we will see more and more automation but also the ability for more powerful analysis. One of the biggest challenges would be to ensure we continue to apply innovative thinking and generate new ideas of how to analyse this data to solve ever-evolving business problems.

In Summary

  • Data Science tools can extract valuable insights to give firms competitive advantages.
  • Data Science can help create early warning systems and be part of risk management.
  • Data Science is continually becoming easier, cheaper and more effective to learn, develop, apply and implement.
  • The expansion and accessibility of Big Data are opening even more opportunities in this space.
  • Actuaries are perfectly suited to embrace the benefits and challenges of Data Science.

Article Contributors

Valerie du Preez (FIA), Dupro Advisory
Michael Jordan (FASSA/CERA), Dupro Advisory

Will Skertic, Data Scientist Beazley
Nick Deveney, Director of Consulting Eden Smith

To learn more about the benefits of Data Science and its opportunities, as well as how to get started, contact