Review of Data Science Solutions to Enhance Pension Communication Theme Conference

Infrared eye-tracking can be used to fairly accurately track how people are reading information on a website. Rein Cozijn (TiU) explains how various eye-tracking technologies are being used in research to collect information on such variables as the frequency with which information is viewed, the “scan path” traversed, the intensity and synchrony of the movements, and what the gaze fixes upon. Dutch insurer NN is investigating the extent to which this technology can provide insight into what and where information should be organized on a page to make the available options clearer and more easily understood for clients and employees.

Watch the presentation by Rein Cozijn (TiU).

Understandability of Informatio
Text mining offers the potential for structuring, analyzing, and categorizing text on a huge scale using vast amounts of data, which can greatly enhance the effectiveness of written communications.  The technique can be used, for instance, to identify the best words to use for reaching certain groups of people. It does require a great deal of data, however, to develop a self-learning algorithm, and the amount of Dutch-language data available is limited. Meanwhile, the Dutch Authority for the Financial Markets (AFM) is faced with the challenge of preventing any given group of the Dutch population from having an inadequate pension in the future. Information plays a critical role in this, and science can help in terms of adapting data and developing models for communication to be used, for instance, in analyzing the understandability of the information presented in tools such as Pension 1,2,3 and improving it. Another important question is how to effectively convey that information to people.

Watch the presentation by Drew Hendrickson (TiU)

Sentiment as a Predictive Quality
Customer communications are increasingly incorporating emotions as a factor. Voice analysis is a reliable method for detecting the affect in human interactions, with textual analysis of that spoken content also important. However, as researcher Marie Postma (TiU) warns, this is certainly not the only variable in predicting customer behavior. Interpreting emotions depends highly on the context of the voice samples being analyzed. In studying a customer service interaction, for example, it might not be readily apparent who is the customer and who the employee. One of the challenges for Dutch insurer a.s.r. has been reaching clients who, while reasonably satisfied with their service, feel that they are not being suitably helped. Speech analysis can be used to reveal potential ambiguities so that a call center employee can pick up on that and provide the right information.

Watch the presentation by Hanka van Waas (a.s.r.).