Anonymisation methods for data from longitudinal studies: BRIDGE funds new research project

28.04.2023 With the exponential growth of digitalisation, ever more personal data is being collected and processed. This data holds great potential for research and industry alike. In a BRIDGE project, researchers from BFH and FHNW are developing AI-based procedures that enable longitudinal data to be made available to research and industry in an anonymised form.

In this, the age of digital transformation, data is being collected electronically in almost all areas of our lives: tracking data from social media apps, movement data, health data or energy consumption data measured by smart meters. In the field of research and development, the huge volume and diversity of data has the potential to create considerable added value. Data protection laws, however, place severe restrictions on the use of personal data. A lot of research has already been done in the past on the anonymised use of data from cross-sectional studies – empirical studies carried out on a one-off basis – with the result that suitable software solutions exist for this purpose. For longitudinal studies – where the same survey is conducted at multiple different points in time – there is still no satisfactory solution that is able to present the data in an anonymised form for research purposes.

Development of anonymisation methods

This challenge is being tackled in a joint project by Murat Sariyar from the Institute for Medical Informatics I4MI at Bern University of Applied Sciences BFH and Matthias Templ from the University of Applied Sciences Northwestern Switzerland. The project is funded by BRIDGE, a programme of Innosuisse and the Swiss National Science Foundation. Core goals of the project are the creation and further development of anonymisation methods that take into account complex data with longitudinal information. The researchers are developing the method using data from the health sector as well as movement data in an industrial context.

teaser Anonymisierungsmethoden für Daten aus Längsschnittstudien

Find out more