By  2 years ago

News media has always been the center of attention regarding the subjectivity of their news reporting. Biasness in news media is of immense interest to various individuals, as the systematic preference of an entity has the potential to invoke support and public action for the benefits of this entity. Although these inclinations are apparent to viewers/followers of the media source, they are still a hindrance to true facts and hence not welcomed. The identification and quantification of media bias is the most sort for metric these days may it be advertising agencies, media regulation authorities or the general public.
In this paper, we present a principled approach to media bias quantification and insightful visualization for popular media sources through their tweets on Twitter. We model the sources and entities of interest by a mini-world of NxN matrix where the count of their tweets and respective polarities over a specified time period are the values. Direct comparisons of the tweet counts and respective polarities will not be meaningful as they do not offer much insight and accurate comparison due to the lack of inherent characteristics of different sources (Twitter participation, the sentimental norm of reporting, etc.) and entities (Popularity, image perception, etc.). Thus, we define coverage and statement scores as properly normalized measures of tweet counts and polarity rates. Furthermore, we present a statistically consistent model of neutral tweet counts and polarity rates using which we define the absolute coverage and statement bias of each source-entity pair. We illustrate our approach on two data sets capturing tweets on 1) Chair person of top political parties of Pakistan in the 2018 general election 2) Paris and Beirut Bombings in 2015 by different sources. Our results demonstrate comprehensive and highly meaningful results. In particular, our proposed lucid visualizations of absolute and relative bias provide a statistically consistent comparison of different media sources on different entities and ready to use results for media managers.