Francesco Grossetti, an Astrophysicist at the Department of AccountingTHE ASSISTANT PROFESSOR DEALS WITH HIGH DIMENSIONAL AND UNSTRUCTURED DATASETS, IN ORDER TO EXTRACT INFORMATION FROM MILLIONS OF PAGES OF FINANCIAL REPORTING AND CODES OF CONDUCT
If linearity bores you, you'll love Francesco Grossetti. Assistant Professor at the Department of Accounting since the beginning of this academic year, Grossetti, before coming to Bocconi, published scientific articles in journals such as Astronomy & Astrophysics or the Journal of Sport Rehabilitation, reflecting an academic path full of twists and turns.
A graduate in Astrophysics and Physics of Space at Milan’s Università Bicocca, he began to tinker with quantitative methods for marketing and management when working for the Politecnico di Milano’s Management and Production Engineering program. «I got close to applied statistics and obtained a doctorate from the Politecnico’s Department of Mathematics, with a dissertation in biostatistics, focused on the use of big data to develop models of hospital compliance to the recommended therapy for patients with heart failure. In practice, I developed an algorithm to make a very high dimensional database treatable on – if I can say - normal computers».
At that time, Miles Gietzmann, the Director of Bocconi's Department of Accounting, was looking for a figure with a solid mathematical basis and capable of developing algorithms to give quantitative robustness to research on Natural Language Processing (NLP), used to extract relevant information from financial reports, codes of conduct and social media posts.
In two working papers of this line of research, Grossetti, through NLP techniques, managed to evaluate the relevance of the different vocabularies used in the evaluation of the tone of a written communication and developed an algorithm capable of determining the optimal number of topics to use when analyzing communications through topic modeling.
Presently, together with colleagues from other departments, he is using NLP and artificial intelligence techniques to extract information from social media posts.
Grossetti is in charge of the Introduction to Blockchain course for bachelor students and supports Gabriel Pereira Pundrich in the Big data for Business Decisions course. He has just published, with Bocconi Professor Gaia Rubera, Python for non-Pythonians, a book aimed at allowing business oriented people to start using the coding language Python.
Find out more:
G. Gavazzi, M. Fumagalli, M. Fossati, V. Galardo, F. Grossetti, A. Boselli, R. Giovanelli, M.P. Haynes (2013): Hα3: an Hα imaging survey of HI selected galaxies from ALFALFA. II. Star formation properties of galaxies in the Virgo cluster and surroundings - Astronomy & Astrophysics.
F. Grossetti, F. Ieva, A.M. Paganoni (2017): A multistate approach to hospital readmissions of patients affected by chronic heart failure: the value added by administrative data - Health Care Management Science.
M. Gietzmann, F. Grossetti, C. Lewis, G. Pündrich: Diagnostics for Textual Analysis of Financial Narrative.
C. Lewis and F. Grossetti: A Heuristic Algorithm for Optimal Topic Identification in Latent Dirichlet Allocation Models.
by Fabio Todesco