Year: 2025
Type: conference-paper
Abstract: Effective workforce analysis, planning, and management require a deep understanding of the tasks and skills associated with different roles. This paper introduces a novel methodology for developing a comprehensive task framework that leverages Large Language Models (LLMs) and large-scale job ads data. We first propose an innovative approach to task taxonomy design, which involves the decomposition and reconstruction of tasks into a hierarchical structure based on action-object pairings, systematically refined using LLMs. The methodology extends to integrating the taxonomy with occupation and skill linkages derived from job ads, ensuring alignment with real-world workforce dynamics. Finally, we demonstrate the practical value of this framework through a visual analytics system that enables interactive exploration and analysis of tasks, occupations, and associated skills, highlighting its potential to transform workforce analysis. Demo video: https://bit.ly/41txBZK
Authors: Rebecca Dew, Mingzhao Li, Weidong Liu, Sandya Baratha Raj
Language: English
Topic: AI and HR Technologies
Domain: Social Sciences
Open Access status: gold
OpenAlex ID: W4410636972