2 Interviews with Actuaries

A major element of the discovery period involved interviewing several actuaries. These meetings served as a key component in both refining ideas for the project, as well as furthering understanding of the industry as a whole. Building on the information from the “R for Actuaries” talks, the interview process allowed the project to narrow its focus and move towards a more well-defined objective.

The project sought to question a broad group of actuaries from a variety of backgrounds. All of the interviewees had considerable experience in either the pensions or life insurance industry. As a result, several key questions could be analysed at a variety of technical levels and perspectives. The various actuaries were also able to provide a sense of what specific areas of their discipline might benefit most from the use of R.

The interview process was particularly insightful with regard to the general use of R and its place within organisations. One of the interviewees had been involved in software selection at a major insurance company. Despite pushing for the use of S-PLUS (a commercial version of R), the company ultimately decided to go with Algorithmics - a well-known and significantly more expensive alternative. A common theme among all respondents was large companies preferring to use established commercial software. As one interviewee put it: “No one ever got fired for hiring IBM.” This attitude provided an interesting perspective into decision-making in the corporate world. The limited adoption of R in large companies is a result of open source and collaborative software lacking the credibility and accountability of a more commercial alternative.

A great emphasis was placed on understanding each of the interviewees’ industries and identifying potential areas that would lend themselves to analysis in R. There was no shortage of ideas, indicating just how applicable statistical tools such as R are to the pensions and life sectors. Of interest were the suggestions relating to predictive underwriting and stochastic mortality estimation. No immediate decisions were made but the respective merits of all proposed ideas were fervently discussed.

An insightful takeaway from the interviews was that the actuaries were particularly impressed by R’s data visualisation tools. Many had been unaware of this strength of R and could see the potential benefits. This raised an interesting question regarding how the project might demonstrate such abilities.

A potential solution came in the form of the R Shiny package [3]. R Shiny is a powerful web framework that allows users to easily create web applications using R. Almost all of R’s extensive data analysis features can be included in this app, which can then be deployed on a hosting service. It was proposed that a functional display of R’s capabilities would serve as a useful guide for current actuaries. Creating such an app became the focus of the project.