Dykstra was the head of ODISSEI during its early years, and the insititute’s innovations in the field of infrastructure are now being used by many scientists. ODISSEI stands for Open Data Infrastructure for Social Science and Economic Innovations. Using funding from the Dutch Research Council (NWO), it set up comprehensive infrastructure for the social sciences over five-year period. Around 45 university faculties, public policy and research institutions and research organisations are members of ODISSEI and pay an annual financial contribution. Using this membership structure, between 80 to 100 people are affiliated with or work directly for ODISSEI. CBS, with the wealth of data that it has at its disposal, is one of ODISSEI’s most important members.
Making data accessible
ODISSEI makes data available to science for research purposes under certain conditions. Using the contributions it receives from members, ODISSEI is able to play a coordinating role and ensure that members work together to make data more accessible, provide support for data collection techniques and share knowledge about algorithms. The result is a growing, thriving community. ‘It has taken years to get people and organisations to work together,’ says Dykstra. ‘ODISSEI is a community that enables research that you could never have dreamed of previously. Now we have to ensure that ODISSEI continues serving the entire research community: from novice researchers to PhD candidates and students.’
What has ODISSEI achieved?
According to Dykstra, the fact that ODISSEI is a shared organisation where people in the social sciences can work together is an achievement in itself. ‘The Dutch Research Council wants to use the ODISSEI model in other areas where there is currently less collaboration. The supercomputer is another great success.’ ODISSEI has been working to extend CBS’s Microdata environment to the SURF supercomputer on a project-by-project basis. This enables significantly faster processing of large data files that require a lot of computing power. Dykstra mentions one example: 'One researcher was analysing segregation and inequality based on a geographical grid of squares measuring 100 metres by 100 metres. Previously, that analysis would have taken months, but using the supercomputer it was done within a few hours.’
Good news for researchers
‘A nice win for science was the accelerated availability of the Long-term Internet Studies for the Social Sciences (LISS) panel during the coronavirus pandemic. The LISS panel asks around 5 thousand households across the Netherlands about savings, sports, health care, education, asylum policy, and so on. Researchers were able to collect real-time data. Once the data was available, it took only six weeks for it to be released.’ The ODISSEI metadata portal is also important, where metadata from various research files is stored in one environment. This makes the metadata more accessible and easier to find. Major suppliers of metadata include CBS, DANS and the LISS panel.
Future developments
As for the future, Dykstra singles out AI: ‘It’s an exciting development. People in the social sciences are experimenting with it more and more, and there’s a lot of interest from adjacent fields, such as life sciences and certain fields of psychology. One concern I have is that AI consumes so much energy. We need to make sure that data storage and AI analysis are less of a drain on our electricity grid.’ Within ODISSEI, there is also a Social Data and Analytics Team (SoDa) at the University of Utrecht. ‘That team is working mainly on AI technology, data collection and data analysis. They are training ever more researchers with new skills. A great asset for the future!.’
