Prof. Wessel Kraaij was inaugurated as Professor of Applied Data Analytics at Leiden University on 24 February 2017. He also hold the post of Principal Scientist at TNO. In both positions, his work focuses on the use of data science. “In areas such as health, the systematic collection of everyday data can create added value for individuals and for society at large.”
“Step by step, our society is becoming increasingly digitized”, says Prof. Wessel Kraaij. “Technological breakthroughs are improving our ability to analyse large amounts of data, to discover previously unknown relationships.” Prof. Kraaij’s research focuses on developing algorithms that can be used to automatically interpret massive amounts of unstructured information. That could involve analysing and describing the content of unknown datasets, or accounting for certain observations or predictions.
Added societal value
According to Prof. Kraaij, if detailed data about a given individual and the environment in which they live is collected and analysed, and then compared to data obtained from a highly specific reference group, this will have genuine value, both for the individual and for the group in question. “It is my belief that providing responsible access to health and treatment information can help enhance learning ability in the healthcare system. Gathering longitudinal data (data collection over a protracted period of time, Ed.) from a large population can be extremely useful, in terms of providing more personal, better health advice, for example. Aside from having a beneficial impact on people’s quality of life, it can also cut healthcare costs”, Prof. Kraaij explains. His research work, both in Leiden and at TNO, focuses on the development of data analytic techniques that create societal value. He points out that, “At TNO, I focus mainly on added value within the healthcare system, in the context of the SWELL and Prana Data projects, for example.”
“Gathering longitudinal data from a large population can be extremely useful, in terms of providing more personal, better health advice, for example – cutting healthcare costs as a result”
SWELL: improving vitality
SWELL, a public-private project which ran from 2011 to 2016, was part of COMMIT/, a national IT research programme. It involved the development of new techniques for measuring stress and effort, and for behaviour-based coaching. “This joint venture between TNO, Radboud University (Nijmegen), the University of Twente, and a number of companies, including Philips, explored the potential use of data science to improve the vitality of employees. The project’s main goal is to develop a digital alter ego that monitors your levels of activity, fitness, and fatigue. By recording activities, social interactions, and moments of concentration, as well as the subject’s emotional, physical and mental condition in a digital diary, or ‘life log’, people can gain a good understanding of their personal performance. These records can be used to spotlight habits that have an adverse impact. This information can then be used to generate personalized feedback which, in turn, can help people to change their behaviour. One of the project’s findings was that facial expressions and posture are potential indicators of stress. As such, these could offer an alternative to determining stress levels by measuring cortisol (a hormone) concentration. But more research is needed.” Prof. Kraaij emphasizes that, “Here, data science does not operate in isolation. When it comes to behavioural changes, the social sciences have a vital part to play.”
“One of the project’s findings was that facial expressions and posture are potential indicators of stress”
Prana Data: pilot projects involving data encryption methods
When recording personal data, it is important to pay special attention to secure storage. “While combining data can generate a great deal of new knowledge, there is also some concern about the security of personal data.” Prof. Kraaij feels that, “Effective data governance is needed, to ensure that members of the public will have no concerns about making their data available.” The pilot projects in Prana Data (a sub-project of COMMIT/) feature data-encryption methods that support simple forms of data analysis. Prof. Kraaij notes that, “We find ourselves in a situation where the technology involved is not yet fully developed and where a range of different ideas are being tested. One of the methods currently being explored involves ‘personal health train’ infrastructure. This makes data available for analysis, even though it may be stored in many different places. Another approach is to use homomorphic encryption (a special encryption method), which protects data, while at the same time supporting simple analyses. In Rotterdam, the ‘Mijn Data Onze Gezondheid’’ (My Data Our Health) trial, which involves the management of personal health data, is currently in the preparatory stages. Pilot projects like this offer practical lessons in how to make the best use of techniques designed to reduce the privacy risks inherent to data analysis. This type of research is still in its infancy.”
“Effective data governance is needed, so members of the public will have no concerns about making their data available”
The idea is that, by exploiting new types of data (such as data from wearables) and by linking datasets together, it will be possible to provide improved personal prognoses and advice. Research is needed to quantify the health gains and the savings in healthcare costs delivered by this approach in practice. “This goal poses a multidisciplinary challenge. In addition, as I mentioned, there are a number of challenges associated with the secure sharing of data (including health data). As yet, however, we do not have the infrastructure needed to manage and share this data. I expect that, in combination with the fundamental research being conducted at Leiden University, the applied research taking place at TNO will contribute to this”, concludes Prof. Kraaij.