Project completed as part of the coursework for the Master's in Data Science at the Barcelona School of Economics. Course: 22DM014 Introduction to Text Mining and Natural Language Processing
It is a common anecdote that books can serve as a source of refuge. Furthermore, there is ample evidence suggesting the benefit of reading fiction in aiding stress relief and even as a coping mech- anism when needing an escape from difficult situations. As we will see in this paper, fiction stories can take many forms. However, some of the best works of fiction are the ones where readers can relate and connect deeply with the emotional experience produced by the plot of the book. These ideas lead us to want to understand if there is a relationship between emotional shocks to society and common emotional arcs found in the books published around the same time as those shocks. To test this, we use World War I (WWI) as an example emotional shock to society and compare the frequency of the emotional arcs shapes in the books published leading up to the war and after it
Along with my teammates Eric Frey and Erika Gutierrez, we examined the underlying story structure in fiction stories in the public domain, (Gutenberg.com). Our goal was to retrieve the underlying story arcs that define different stories, and try and correlate the distribution of these arcs before and after the world war to see if there is a statistically significant difference to be found.