In this project, I examined two works of detective fiction using digital techniques.  I first analyzed The Name of the Rose by Umberto Eco using character-specific techniques in order to understand how Eco portrays suspects and victims.  I wrote programs in the R programming language in order to extract pieces of text that mention a given character’s name.  I then created a dispersion chart by piecing together all the passages that include a character’s name to create that individual’s timeline within the text.  Plotting mentions of a second character in that timeline revealed the frequency with which the second character appears in relation to the first.  Additionally, I used Google’s topic modeling tool with these selective passages to see how Eco most frequently describes the characters. I found that Eco typically mentions the murderer and the victim together around the time of the victim’s death.  Additionally, he often uses words related to the method of murder in close proximity with the murderer’s name.  Following this analysis, I repeated the procedure with Agatha Christie’s The Mysterious Affairs at Styles in an attempt to predict the plot of the novel.  This prediction was incorrect, likely due to differing naming practices.  While I used the name “Howard” to identify the character Evelyn Howard, Christie more frequently refers to her as Evie, which likely altered my results.  Without a corrected naming schematic, it is unclear whether the results of The Name of the Rose also apply to The Mysterious Affairs at Styles.

Phoebe Bumsted

Bowdoin College