This thesis applies text mining techniques to analyze the thematic content of Supreme Court opinions and the voting behavior of justices between 2000 and 2020. Latent Dirichlet Allocation (LDA) revealed five major legal themes involving Tribal Sovereignty and Land Rights, Criminal Procedure and Punishment, Education and Religion, Employment and Discrimination, and Elections and Political Speech. A chi-squared test further showed that justices appointed by presidents of the same political party were significantly more likely to vote together. The results highlight partisan coalitions and cross-party alignments, demonstrating the value of text analysis and statistical methods to better understand judicial decision-making.