Modeling Literary Preferences Using Complex Networks and Centrality Measures

For analyzing the preferences of consumers for certain types of literature, descriptive statistics is extensively used by both researchers and book readers. However, using only descriptive statistics methodology to interpret the individuals’ preferences to read novels, poems, short stories, theater, or other book types is a limited approach. We suggest in this work using both descriptive statistics and node centrality measures of complex networks in order to better identify the books with influence over the preferences of a group of individuals. Using the software RStudio, a weighted undirected network is designed for the application of the study, by representing a group of individuals, their preferred books, and the preference intensity. We present results and conclusions about representing individuals preferences using network structures, and about measuring most influential network nodes through centrality indicators: degree, closeness, betweenness, and eigenvector centrality. Given that the book is a cornerstone of a civilized and educated society, we believe it should be more present among research topics, in order to encourage people (particularly young people) to read.