Rethinking Data Science (ReDS): advances in causal inference and statistical learning across domains – DiSIA Project of Excellence 2023-27

Five intense years await the Department of Statistics, Computer Science, Applications “Giuseppe Parenti” (DiSIA) , once again selected among the “Departments of Excellence”. The recognition by the Ministry of University and Research – and the consequent request for funding by the DISIA of approximately 5.4 million euros – are due to the strategic development project ” Rethinking Data Science: advances in causal inference and statistical learning across domains ( ReDS ) ”.

DiSIA has become a relevant national and international center for the disciplines involved in the process of knowledge production through the intensive use of data to support decision-making processes. However, recent developments in Data Science highlight the limits of some analytical tools and algorithmic approaches that learn from a large amount of data with mainly predictive purposes. The limited attention to data quality, difficulties related to replicability and generalizability, as well as the lack of causal interpretation, cut down the usefulness of these tools in decision support.

With ReDS, the 2023-2027 project of excellence, the DiSIA aims to overcome these critical issues by giving a further boost to its research and teaching activity; this project will lead the department to become an international hub for the statistical and causal analysis of complex data as well as for the training of a new generation of Data Scientists.

Carla Rampichini, director of the department declared to UnifiMagazine, the official magazine of the University of Florence:

“We are very happy with this success. A confirmation both of the excellent work done in the last five years, given that we received a better score in the Evaluation of Research Quality compared to the last selection, and of the considerable effort in drafting the new project. Not to mention her – she adds – that in the context of funding destined for the economic sciences and statistics area, at a national level only our department deals specifically with statistics and information technology. The incoming funds will give further impetus to the growth of DISIA, with a view to a future self-financing capacity”.

The main research goal is the development of statistical methods for complex data and for causal inference that can be used as decision-support tools in specific domains, combining good forecasting value with interpretability, while clearly quantifying the uncertainty. These methods will provide comprehensive support to scientific knowledge in the DiSIA domains of interest as demography, social sciences, economic phenomena, medicine and epidemiology.

The main educational goal is the strengthening of the department’s undergraduate and graduate programs and the training of excellent data science researchers. ReDS aims at the transfer of knowledge and professional training by promoting and stimulating interactions between DiSIA researchers and people interested in the project themes, both within and outside academia.

Fabrizia Mealli, scientific director of the project as well as director of the Florence center for Data Science, in the interview by UnifiMagazine,explains:

Research in Statistics, Statistical Learning, Machine Learning and Artificial Intelligence are galloping and also the methodological activity based on data analysis has taken important steps in these five years. Various critical issues have been highlighted regarding the use of data without a precise structure and collected without a particular design. It clearly emerged that these shortcomings should be addressed by recovering notions of probability and statistical inference.

Quantifying uncertainty adequately will support Machine Learning and Artificial Intelligence algorithms, very useful in making forecasts, but still imperfect in quantifying the uncertainty of these forecasts. With ReDS, DISIA intends to strengthen its contribution in implementing this approach to data analysis for both scientific discovery and decision support purposes.

DiSIA has been selected for the second consecutive time as department of excellence, the first one being the Project of Excellence 2018-2022

Read the full interview on UnifiMagazine to Carla Rampichini, Head of DiSIA and Fabrizia Mealli, scientific coordinator of the project (Italian version only).