Public Interest in Razón Pública: a Data-driven Network Analysis
Abstract
This thesis is a case study based on a network analysis of the magazine Razón Pública. This project took an inductive, data-driven approach to public interest in the context of a specific digital network. The main objectives were to characterize the magazine’s authors and subscribers, identify its prevalent news topics, and validate to what extent subscribers agree that the resulting topics are of public interest or not. In order to achieve these objectives, a mixed methodology was used; it consisted of combining Natural Language Processing (NLP) techniques -keyword network, n-grams analysis, and Latent Dirichlet Allocation (LDA) topic modeling-, qualitative analysis of a sample of articles, and surveys. It was found out in this research that men and people with higher education levels were overrepresented among Razón Publica’s authors and subscribers; it also concluded that subscribers, largely, agreed on considering the topics extracted from the magazine as of public interest.