by Maya Peacock (M.S. Data Science)

Maya Peacock used data from the Jane Addams Digital Edition to determine how best to use sentiment analysis tools to interrogate historical documents. Sentiment analysis is a computational process where words are assigned to determine whether words, and the documents that contain them, carry positive or negative aspects. She used two tools, VADER and TextBlob, to determine the best means to study historical texts, looking to find known periods in Jane Addams’s life when we might expect to find more negative or positive content and seeing whether the tools can find the same patterns. Maya found there were some issues with using tools designed to scrape current web content when applied to historical texts, but she was able to locate time periods, such as during World War I, when discussions were more negative. These corresponded to subject tags such as prisons, racism, and war, proving that the network was able to work with the texts. She also investigated whether she could identify people whose letters were more positive or negative, finding a few instances of polarizing people who scored highly on the scales.