DTA

Digital Theses Archive

 

Tesi etd-02142023-223310

Type of thesis
Corso Ordinario Secondo Livello
Author
TESTA, LORENZO
URN
etd-02142023-223310
Title
Unveiling research patterns of institutions via networks and functional data: evidence from genome research
Structure
Cl. Sc. Sociali - Scienze Economiche
Course
SCIENZE ECONOMICHE E MANAGERIALI - SCIENZE ECONOMICHE E MANAGERIALI
Committee
relatore Prof. MINA, ANDREA
Relatore Prof.ssa CHIAROMONTE, FRANCESCA
Presidente Prof. IRALDO, FABIO
Membro Prof.ssa ANNUNZIATA, ELEONORA
Membro Prof. PICCALUGA, ANDREA MARIO CUORE
Membro Prof.ssa MAGAZZINI, LAURA
Membro Prof. MOSCHELLA, DANIELE
Membro Prof.ssa ROMANO, MARIA FRANCESCA
Membro Dott. GIACHINI, DANIELE
Keywords
  • bibliometrics
  • functional data analysis
  • genome research
  • network science
  • Science of science
Exam session start date
20/03/2023;
Availability
parziale
Abstract
Scientific productivity follows trajectories, shaped by heterogeneous factors that are not always easy to capture. In this work, we analyze bibliographic data on genome research and introduce state-of-the-art methodologies which shed new light on scientific productivity, describing some of its patterns. We retrieve data on publications from OpenAlex, collecting -- among the others -- information on authors' affiliations. Then, by exploiting our data, we build a collaboration network among institutions. We dynamically extract from it numerous measures of centrality, which heterogeneously embed the underlying network topology. We treat these dynamic observations as curves, and we exploit tools from functional data analysis to analyze them. We perform functional clustering on the institutional productivity trajectories, highlighting the strong geographic patterns behind them. Then, we run fully functional function-on-function regression models in order to link collaboration measures to productivity, and to quantitatively measure their impacts. We discover that institutions collaborating more (in particular in terms of acting as bridges among collaborators -- a feature measured through the betweenness centrality) are also more productive, in terms of number of papers published across time.
Files