PREDICTIVE ANALYTICS FOR EMPLOYABILITY IN CREATIVE FIELDS

The field of employability assessment is getting revolutionized by predictive analytics in creative industries because it can be used to assess skills, portfolios, and career paths based on data. The creative labor markets (design, visual arts, media, and digital content production) can be described by their heterogeneous skills, subjective quality standards, and quickly changing demand trends, which renders the conventional methods of employability assessment ineffective. This paper hypothesizes a predictive analytics system to be used in forecasting employability within the creative industry by incorporating portfolio artifacts, talent profiles, project assessments, and labor market indicators. The methodology is a mix of structured attributes based on education, experience, and performance metrics and unstructured textual and visual data taken out of portfolios and job adverts. Various predictive models are used such as Random Forest, XGBoost, Artificial neural networks, and BERT-based language models to align the nonlinear relationship and semantic alignment between creative work and market need. SHAP-based explainability helps to determine which drivers of employability, including skill diversity, project relevance, aesthetic coherence, and industry alignment are most important using model interpretability. The results of the experiments prove that ensemble and deep learning models are superior to traditional ones, they are more accurate and robust in employability prediction in creative sub-sectors. This can be used in practice to create AI-based portfolio analysis, a customized identification of skill-gaps, and integration with online talent markets to enhance workforce planning and development.