This study introduces a novel method for detecting and visualizing temporal trends in topics, moving beyond traditional stochastic counting to illustrate both the significance and directional movement of topics as defined by subtopics within the corpus. Utilizing k-means clustering and cosine similarity (derived from word embeddings in a vector space), this method maps the convergence or divergence of topic clusters over time. Furthermore, when applied to a diverse corpus that includes media articles from our dataset, our approach uncovers how topics merge or become distinct within a text corpus by analyzing the convergence and divergence of embeddings in a vector space, providing deep insights into public discourse and information flow during major global events. This methodology not only enhances understanding of health-related topics like COVID-19-capturing public sentiment and the dynamics of information dissemination-but also shows promise for broader applications across various general topics, offering a powerful tool for analyzing temporal topic evolution in any given field.