This study analyzes historical speeches from the United Nations General Debate, spanning multiple decades, to uncover trends in topics, and language usage. Utilizing a dataset of over 4,700 speeches from various countries and years, we apply natural language processing (NLP) techniques in Python to process and visualize the data. Key methods include tokenization with NLTK and spaCy for text cleaning, word frequency counting for visualizations like word clouds and histograms, and topic modeling via Latent Dirichlet Allocation (LDA) and Non-Negative Matrix Factorization (NMF) to identify dominant topics such as international peace, economic development, and conflict resolution. The analysis reveals evolving patterns, including a shift toward climate and inequality discussions in recent years and regional variations in language focus. Visualizations highlight topic proportions over time, showing increasing emphasis on sustainability post-2000, as well as the different issues of interest for different countries. These insights demonstrate the impact of geopolitical changes on diplomatic discourse and underscore the role of data science in extracting meaningful patterns from large text corpora. The findings provide valuable guidance for policymakers, researchers, and international relations experts in understanding global priorities and fostering data-driven diplomacy.