A Multi-Faceted Evaluation Framework for Assessing Synthetic Data Generated by Large Language Models