In the Data Exchange project, we are investigating what role a trusted third party could play. We also want to find out what checks and balances are required so that a data provider can trust that the data is only used in the manner that was agreed upon.
Prototype of a platform
Data providers such as companies or academic hospitals want to make their data available for research, but want to control who uses the data and for what purpose. They also have to comply with legal requirements in the field of personal data. SURF has built a prototype of a platform on which data can be shared without losing control over the confidentiality of the data.
How does that work? A researcher requests a data provider to run a specific analysis on a specific dataset. The data provider reviews the request, and after approval, the analysis is run by a third party in a secure container, ensuring that no data is leaked to the outside world. Once the analysis is completed, the data provider can curate the results before deciding to release them to the researcher. These checks ensure that no unintended data is ever released, while it allows researchers to perform analyses on confidential datasets.
To alleviate the burden of manual verification for each calculation that is performed, it is possible to allow a specific analysis on multiple datasets, or to allow a researcher to run any analysis on one specific dataset. These permissions can be withdrawn at any time.
Webinar Sharing confidential data for research - 23 June 2020
During this webinar we presented 4 projects in which research is done on confidential data, while the data suppliers maintain control over their data.
- Nanda Piersma (HvA and CWI): short keynote
- Ruurd Schoonhoven (CBS): CBS Microdata Services
- Ananya Choudhury (Maastro): Personal Health Train
- Leon Gommans (UvA/Air France KLM Group): Prototype of a Digital Data Marketplace
- Freek Dijkstra (SURF): SURF Data Exchange
- Poll with the audience
Watch the recording below:
What do you think?
Using this prototype, we want to ask both data providers and researchers what guarantees are needed, in what circumstances, to make it easier for researchers to use confidential data. We also want to find out how such an infrastructure fits in with existing data workflows.
Are you interested in this topic? We would very much like to get in touch with you to exchange ideas and learn how such a concept could help you.
This project is part of the SURF Open Innovation Lab.
Project team SURF