SentiAuthor: A Web Application that Displays Indications of Sentiment, and Potential Biases, for Individual Contributors to a News Organization

Document Type

Project

Lead Author Type

CIS Masters Student

Advisors

Dr. Jonathan Leidig, jonathan.leidig@gvsu.edu

Embargo Period

5-8-2015

Abstract

Sentiment analysis (SA) has become a vibrant area of research over the past several years. By and large, SA has been used to extract sentiments from an aggregation of numerous contributors’ articles or reviews about a specific subject. For example, the cumulative evaluations of multiple reviewers are often examined en masse to determine whether customers have a positive or negative feeling toward a product or service. Similarly, the aggregated work of hundreds of journalists may be analyzed to estimate the overall political bias of an entire news organization. However, there are relatively few studies in the SA literature which examine the sentiments, and potential biases, of one particular contributor at a time, by taking into account that author’s body of work over a period of time. To address the lack of exploration in this area, a web application was created which has access to all of the more than 650,000 articles in Guardian News’ database over the past five years, and displays indications of sentiment for a single contributor at a time. Users are able to select a specific author, and view that contributor’s sentiments regarding the themes, entities and aspects that occur most frequently in their most recent work. The ultimate goal of this project is to empower users to utilize the displayed information to draw their own conclusions about the sentiments, and potential biases, that the author may possess.

This document is currently not available here.

Share

COinS