Knowledge Workers and Regional Economic Disparities in Italy: A Data-Driven Analysis

The article analyzes knowledge workers' values across Italian regions. Metric analysis is carried out by using the ISTAT-BES database. To assess whether any regional grouping is present, based on knowledge worker values, the paper applies clustering by using the k-means algorithm and then optimizes it with the Silhouette coefficient. The data show that the optimal number of clusters corresponds to k=3, but this corresponds to a suboptimal value of the Silhouette coefficient. It is followed by the panel data analysis in order to check for the metric relationships between knowledge worker values and a set of innovation-related variables observed in Italian regions. The results show that knowledge worker values increase with the presence of cultural and creative employment and regular internet users, while they decrease with higher household internet access, increased mobility of university graduates, and research intensity. Finally, a machine learning analysis identifies the most efficient algorithm in prediction the future value of knowledge workers by optimizing the metrics R-squared and Mean Absolute Error. Findings also come with a discussion in the context of the North-South divide that characterizes Italian regions. JEL CODE: O3, O31, O32, O33, O34