Visualising Sentiment in the Play The Demon Monk

During her professional placement with the Digital Scholarship team, part of her Master’s programme in Archaeology, Yifan Zheng explored data from our digitised collections. She focused on materials from the Rosicrucian collection, which contains publications of the Rosicrucian Order’s Crotona Fellowship.

Yifan used an API key to access the full text of a play in the collection, saving the text to a file for ease of use offline. She used a Jupyter Notebook to break the text into acts and scenes. Each scene was further broken into sentences that were analysed using the Natural Language Toolkit (NLTK) to obtain sentiment scores. The sentiment scores can be between -1 and 1, a score above 0.05 is positive, a score below -0.05 is negative and a score between -0.05 and 0.05 is neutral. By averaging the sentence scores within each scene, she produced a sentiment score for every scene in the play.

Finally, Yifan visualised these results using the matplotlib Python library, creating a graph that highlights how sentiment shifts from scene to scene throughout the play.