Uncertainty, Data and Trust

How can novel data and methods be used to address urgent problems of our time? This was discussed at the online event.

The interpretation of research data invariably also involves dealing with uncertainties. What is trustworthy, what is proven knowledge and what is mere hypothesis?

The event "Uncertainty, Data, Trust", which was broadcast live from the Museum of Communication via Zoom on 10 June, was not about the very topical issue of the corona pandemic, but rather about climate change. Marius Zumwald and Benedikt Knüsel, who completed their doctorates under climate researcher Reto Knutti as part of the NRP 75 project "Can a combination of theory and Big Data better predict extreme weather impacts?", spoke about their research projects.

What makes predictions trustworthy?

For a study on heat in cities, for example, Marius Zumwald supplemented the "classic" data from MeteoSwiss measuring stations with innovative data from private weather stations operated by urban dwellers. Thanks to Big Data and machine learning, he was able to create a detailed model of heat distribution in the city. Benedikt Knüsel, in his short presentation, dealt with questions such as what exactly makes model predictions trustworthy. In addition, he illustrated what is meant by "accuracy".

The discussion, which also engaged the audience, focused on issues of trustworthiness of data and models, as well as the conclusions that can be drawn from them.

This interactive event was a co-production of the Museum of Communication, the National Research Programme NRP 75 "Big Data" and the Science et Cité Foundation.