Document Type

Article

Publication Date

6-2026

Publication Title

Socio-Economic Planning Sciences

Abstract

The digital transformation and the subsequent datafication of work generate rich electronic traces that can be transformed into actionable insights through modern analytics. In this context, this study proposes a data-driven framework that leverages sidClustering and unsupervised Random Forest (RF) for clustering and feature selection, constructs composite indicators via multiple aggregation strategies, and employs visual tools to enhance the interpretability of results. A key feature of the framework is the integration of two data sources: click metadata from Microsoft 365 and employee attitudes measured through a questionnaire. This integration enables the analysis of digital work behavior (DWB) in relation to employee sentiment. We apply the framework to a highly digitalized Italian consulting company. The analysis identifies two employee clusters, 'Operational' and 'Coordination', and yields two synthetic digital work metrics, work Quantity and Complexity. Overall, the study introduces a scalable methodological framework that combines tree-based learning, composite indicators, and visual tools, representing one of the first empirical integrations of digital work metadata with employee attitudes. The resulting indicators offer early-warning capabilities for assessing the impact of technology adoption on employee outcomes and provide decision support for HR analytics and policy.

Comments

Open access Funding provided by University of Modena and Reggio Emilia within the CRUI-CARE Agreement.

DOI

10.1016/j.seps.2026.102496

Version

Publisher's PDF

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

Download

Included in

Mathematics Commons

Share

COinS