Data Science education is experiencing annual enrollment growth, driven in part by government initiatives aimed at promoting racial equity in STEM fields. It is vital to ensure that all students have access to the necessary learning resources. While gender bias in STEM has received significant attention, research on racial bias is relatively limited, and the combined effects of gender and racial biases in Data Science remain largely unexplored. The objective of this study is to investigate these biases in Data Science education by examining the preferences of 300 diverse students enrolled in Data Science related classes, with a specific focus on their preferences regarding teaching practices and methods. The impact of race and gender in Data Science education is an important area of study that has gained increased attention in recent years. Race and gender play a critical role in representation within the Data Science community. Stereotypes and biases can influence the experiences of individuals from different racial and gender backgrounds within Data Science education. Race and gender can impact data science pedagogy in various ways, influencing instructional practices, content delivery, and student experiences. Using machine learning techniques to understand biased pedagogy in data science education can be a valuable approach to uncover patterns and insights. We propose an effective unsupervised learning approach to examine the correlations between preferred pedagogical methods and gender-racial background of the participating students. Our findings show that student's gender has a significant correlation with learning style; e.g., female students prefer more private methods of interaction with TAs and instructors, such as anonymous polls and Zoom meetings. Furthermore, our study uncovers significant associations between race, gender, and learning styles, showing that Asian students exhibit higher levels of Classroom Interaction and Group Assignment Preference tha...