Description
Machine learning has opened an opportunity to study complex social process using qualitative data in new and rich ways. Using these methods to better understand society does come from reducing complex, qualitative data to numbers, however, but in visualizing data in ways that preserve the complexity and embeddedness of social processes while also illuminating their empirical dynamics. I use Python to implement a word embedding model of first-person narratives of the nineteenth-century U.S. South, using the output from this model to visualize one social process, intersectionality, in a way that is reproducible and stays true to the epistemology of intersectional research. Python, machine learning, and visualizations, I argue, can help explain complex social processes without reducing them to uninterpretable numbers on a page.