skip to content

Turing Gateway to Mathematics

 

Techniques for Data Linkage and Anonymisation
23rd October 2014 - Isaac Newton Institute, Cambridge

Data Linkage

Background

Increasing numbers of large databases describing people, their characteristics, and their behaviour are being looked to for analysis by both Government and commercial organisations. Much can be gained from combining or linking data from different datasets leading, for example, to better understanding of the needs of the public and to better public policy decisions. However, a balance needs to be struck between the need to access data and the need to preserve the privacy and confidentiality of personal information.

Within this context, more research is needed to develop techniques for the safer linkage/merging of data, as well as those for more effective anonymisation in order to preserve privacy and confidentiality of personal information, while enabling analysis which will benefit the public.


Aims and Objectives

The two areas - data linkage and anonymisation - are fundamentally interdisciplinary. With this in mind, this workshop brought together leading experts from a variety of relevant disciplines, including mathematics, computer science and statistics, as well as those from other related disciplines. Likewise, a key objective was to gain input from key stakeholders, with expertise in specific application domains, such as educationalists, official statisticians, medical researchers, commercial organisations, epidemiologists and others.

The workshop examined:

  • how to effectively link or merge multiple databases describing individuals, including how to cope with inaccuracies, partial overlap, ambiguous records, and so on, in a trustworthy way. This is the question of linkage.
  • how to release the information in the databases, merged or not, for research and policy purposes without compromising the privacy of the individuals in the database. This is the question of anonymisation.

The workshop included reviews of current methods and state-of-the-art in the two areas, but also aimed to identify the critical challenges faced in the areas of data linkage and anonymisation.

This event was attended by individuals from a number of areas including:

  • Biomedical and health research
  • Social and economic research
  • Government, policy makers, regulatory authorities and statisticians
  • Commercial organisations in the financial and retail sectors, and others

Venue

Please see the Isaac Newton Institute A-Z for further information about the venue.

Support

This event was supported by ESRC.