Abstract

This research pioneers a predictive approach to assessing unemployment rates and NEET (Not in Education, Employment, or Training) populations in Indonesia through an advanced integration of mobile positioning data (MPD) and content analysis of news articles. Employing a dual-source data model, the study leverages the geographical and behavioural insights gleaned from MPD and the dynamic, real-time socio-economic narratives extracted from extensive media monitoring. The synthesis of MPD enables the identification of mobility patterns and their associations with regional employment statistics, providing a novel predictive layer for unemployment and NEET tendencies. Simultaneously, sentiment and thematic analysis of digital news content serve to predict discrepancies between educational achievements and job market requirements. This dual analytical strategy enhances the traditional use of static datasets with a richer, more textured understanding of labour market dynamics. Predictive models are developed and refined through machine learning techniques, employing algorithms that interpret complex datasets to forecast socio-economic conditions. Validation against official statistics from BPS-Statistics Indonesia ensures the accuracy and reliability of the predictive outcomes. Findings of this research offer actionable intelligence for crafting targeted educational and employment policies. By establishing a predictive nexus between mobility, media narratives, and labour market realities, this study sets the stage for informed policy-making that proactively addresses educational and employment mismatches in Indonesia’s evolving economic landscape.