Digital detection of play characters’ relationships in Shakespeare’s plays: extended cross-correlation analysis of the character appearance frequencies

We propose a method for visualizing literary works that quantitatively extracts the mutual relationships among play characters from the narrative of a storyline. The method first determines the cross-correlation of the appearance frequencies in the time domain between two play characters, which is calculated for all pairs of characters in each narrative. We also calculate the correlation among three play characters to find unique triangular relationships. Then we create a graphical representation of the relationships using node-link representations based on a physical potential model. The method is suitable for dramas, as demonstrated for ten famous Shakespeare plays. The resulting visualizations show good agreement with the conventional understanding of each play and also provide new insight into Shakespearean criticism.