Data science and statistical methods are at the heart of the collaboration between the University of Neuchâtel (UniNE) and the Federal Statistical Office (FSO). Their collaboration agreement 2021-2024, signed at the end of February, follows the creation of a Data Science and Statistical Methods Division at the FSO and a new Chair in Data Science at UniNE. The aim is to promote research and the development of concrete projects in the field of digitisation.

The FSO is currently developing a structure to meet the challenges of the digitisation of data science and statistical methods. A new division has been created under the umbrella of this area. Over the next few months, a centre of competence for data science will be set up to meet the needs of the entire Confederation with services ranging from consulting and training to methodological support and the complete implementation of projects in this field.

For its part, UniNE will create a chair in data science with a focus on artificial intelligence research. The newly hired person will have the task of developing scientific projects of international scope within the Institute of Informatics.

Projects to improve quality and efficiency

Collaboration between the SFO and the UniNE has existed for some twenty years. Within the Institute of Statistics, it has resulted in the publication of around ten doctoral theses and numerous articles in peer-reviewed scientific journals. For the FSO, this partnership has made it possible to set up a system for sharing the burden of surveys at both the enterprise and individual levels, as well as numerous improvements in employment and wage statistics, for example.

The launch of the pilot projects described on the FSO’s “experimental statistics” microsite has enabled a new stage to be reached. These innovations should help to further increase the efficiency of current techniques. For example, the ADELE project (“Arealstatistik Deep Learning”), which uses artificial intelligence techniques in the context of aerial image recognition for national area statistics, clearly illustrates the potential for automation and thus productivity gains.

Increased attractiveness thanks to data

For the UniNE, the great diversity of the topics dealt with at the SFSO, their particular problems and above all the access to real data (secure access and a data protection contract being obviously required in each project) is a source of inspiration which helps to attract a large number of researchers. The staff who collaborate under this agreement, most often doctoral and/or post-doctoral students, thus see the results of their research find direct use in the statistical production of the FSO.

This example of partnership is not isolated. The FSO also collaborates in the field of data science, including artificial intelligence, with other universities and universities of applied sciences in the country as well as with the two Federal Institutes of Technology and their “Swiss Data Science Center”.

A resolutely forward-looking vision

In order to intensify this collaboration, the FSO has launched an internal call for new projects in the future-oriented field of data science and statistical methods. The aim is to cover the numerous topics dealt with by the FSO and to provide an opportunity to develop collaboration in concrete projects such as those relating to codification or the processing of missing and/or outliers.

Translated from Suisse : L’Université de Neuchâtel et l’Office fédéral de la statistique collaborent sur la science des données et les méthodes statistiques