Data-Driven Exploration of Skill Mismatch: Leveraging Textual Analysis for Comprehensive Insights

Ibrahim Rahhal, Mounir Ghogho, Kenza Benchaaboune

Abstract: Skill mismatch can be attributed to diverse factors, encompassing but not limited to outdated or misaligned curricula, swift technological advancements, alterations in economic structures, inadequate training programs, and a dearth of comprehensive information about the job market. Prior investigations predominantly employ survey-based methodologies to acquire data and derive insights. However, this approach often proves time and resource-intensive and encounters limitations related to respondents' knowledge constrained within their specific domains. Moreover, the insights generated are circumscribed to the surveyed area of study and lack generalizability. In this paper, we investigate skill mismatches in cybersecurity education and job market requirements, focusing on the alignment of university curricula with industry demands. We conduct a comprehensive analysis of job advertisements and university training programs to identify disparities in skills, particularly in programming languages, certifications, and soft skills. We use data science techniques, offering a more detailed and current view than traditional surveys. Key findings include mismatches in specific skills like C#, C++, Sql Server and certifications, absent in university training, which highlights the need for curriculum evolution to enhance graduate employability and recommends incorporating in-demand skills to meet the cybersecurity job market's dynamic demands.