ABSTRACT This thesis examines the economic effects of national level digital transformation on regional economic performance across 243 NUTS-2 regions in 27 EU member states from 2015 to 2023. While existing literature analyzes digital transformation predominantly at the country level, this study argues that national conclusions conceal profound regional heterogeneity — the same digital policy produces fundamentally different outcomes across structurally distinct regions within the same country. Using 2,133 region-year observations from Eurostat, this thesis applies descriptive statistics, hypothesis testing, K-Means clustering, hierarchical clustering, multiple linear regression, and logistic regression to examine four digital indicators — cloud computing, artificial intelligence, e-commerce, and digital skills — against regional GDP and unemployment outcomes. Three findings emerge. First, EU regions cluster into three distinct digital typologies — High, Medium, and Low Digital — corresponding to Nordic and Western European, Continental European, and Eastern and Southern European geographic patterns. Second, AI and cloud adoption are the most economically significant indicators producing statistically significant positive effects on GDP and negative effects on unemployment probability. Third, digital skills exhibit a paradoxical negative relationship with economic performance, suggesting a structural skills mismatch where digitally trained workforces lack sufficient business level digital infrastructure to absorb them productively. These findings require national digital strategies to incorporate regional differentiation — treating all regions as economically identical produces misleading conclusions and ineffective policy. Keywords: Digital Transformation, NUTS-2 Regions, K-Means Clustering, Regional Economic Performance, Skills Mismatch, European Union.
RIASSUNTO Questa tesi esamina gli effetti economici della trasformazione digitale a livello nazionale sulla performance economica regionale, analizzando 243 regioni NUTS-2 in 27 stati membri dell’Unione Europea nel periodo 2015-2023. Mentre la letteratura esistente analizza prevalentemente la trasformazione digitale a livello nazionale, questo studio sostiene che le conclusioni aggregate nascondono una profonda eterogeneità regionale — la stessa politica digitale produce risultati fondamentalmente diversi in regioni strutturalmente distinte all’interno dello stesso paese. Utilizzando 2.133 osservazioni regione-anno da Eurostat, questa tesi applica statistiche descrittive, test di ipotesi, clustering K-Means, clustering gerarchico, regressione lineare multipla e regressione logistica per esaminare quattro indicatori digitali — cloud computing, intelligenza artificiale, commercio elettronico e competenze digitali — rispetto al PIL regionale e al tasso di disoccupazione. Emergono tre risultati principali. In primo luogo, le regioni europee si raggruppano in tre tipologie digitali distinte — Alto, Medio e Basso Digitale — corrispondenti ai modelli geografici nordici e dell’Europa occidentale, continentale e meridionale-orientale. In secondo luogo, l’adozione di AI e cloud computing emerge come la più significativa economicamente. In terzo luogo, le competenze digitali mostrano una relazione paradossalmente negativa con la performance economica, suggerendo un disallineamento strutturale delle competenze. Questi risultati richiedono che le strategie nazionali incorporino una differenziazione regionale — trattare tutte le regioni come economicamente identiche produce conclusioni fuorvianti e politiche inefficaci. Parole chiave: Trasformazione Digitale, Regioni NUTS-2, Clustering K-Means, Performance Economica Regionale, Disallineamento delle Competenze, Unione Europea.
DIGITAL TRANSFORMATION AND REGIONAL ECONOMIC PERFORMANCE IN THE EUROPEAN UNION: A MACHINE LEARNING AND ECONOMETRIC ANALYSIS OF NUTS-2 REGIONS
MEKONENE, YOHANNES JIMA
2025/2026
Abstract
ABSTRACT This thesis examines the economic effects of national level digital transformation on regional economic performance across 243 NUTS-2 regions in 27 EU member states from 2015 to 2023. While existing literature analyzes digital transformation predominantly at the country level, this study argues that national conclusions conceal profound regional heterogeneity — the same digital policy produces fundamentally different outcomes across structurally distinct regions within the same country. Using 2,133 region-year observations from Eurostat, this thesis applies descriptive statistics, hypothesis testing, K-Means clustering, hierarchical clustering, multiple linear regression, and logistic regression to examine four digital indicators — cloud computing, artificial intelligence, e-commerce, and digital skills — against regional GDP and unemployment outcomes. Three findings emerge. First, EU regions cluster into three distinct digital typologies — High, Medium, and Low Digital — corresponding to Nordic and Western European, Continental European, and Eastern and Southern European geographic patterns. Second, AI and cloud adoption are the most economically significant indicators producing statistically significant positive effects on GDP and negative effects on unemployment probability. Third, digital skills exhibit a paradoxical negative relationship with economic performance, suggesting a structural skills mismatch where digitally trained workforces lack sufficient business level digital infrastructure to absorb them productively. These findings require national digital strategies to incorporate regional differentiation — treating all regions as economically identical produces misleading conclusions and ineffective policy. Keywords: Digital Transformation, NUTS-2 Regions, K-Means Clustering, Regional Economic Performance, Skills Mismatch, European Union.| File | Dimensione | Formato | |
|---|---|---|---|
|
THESIS (6) (1).pdf
embargo fino al 18/01/2028
Descrizione: Bachelor Thesis - Digital Economics and Business
Dimensione
559.13 kB
Formato
Adobe PDF
|
559.13 kB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.12075/27474