Semiautomatic Generation of Graphics for Data Journalism Using Open Data

A Case Study of Higher Education Census

Abstract

Computational Journalism refers to the tools and algorithms that journalists use to tell stories. In this context, open data can serve as a news source. Although various data manipulation techniques, such as Data Warehouse, Data Mining, and Business Intelligence, can answer questions about data, their responses are often static and noninteractive, limited to predefined queries. This project aimed to develop an application that allows the generation of dynamic graphs from the instant selection of filters applied to data from the Higher Education Census of Brazil. This approach was intended to facilitate the interpretation
of this data by journalists. The implemented prototype was evaluated using the SUS scale and achieved satisfactory results in terms of system usability. As a result, this approach enabled the extraction of new information that can be used in the creation of news articles.