Earthbound Topic Modeling and Sentiment Analysis (WIP)

The goal of this project is to to create a topic model from the text of the game Earthbound, and to run a sentiment analysis on quotes related to various topics to see how sentiments compare between topics.

Credits

The text dump was made by Clyde "Tomato" Mandelin and was sourced from here: https://starmen.net/mother2/gameinfo/script/

Packages used

• NLTK (vader_lexicon, punkt, wordnet, stopwords)
• spacy
• gensim (LDA topic modeling package)
• pyLDAvis (LDA visualization)
• IPython (dependency of pyLDAvis)