Examining Government-Citizen Interactions on Twitter using Visual and Sentiment Analysis

ANA MAGUITMAN; ELSA ESTEVEZ; ROCIO HUBERT; TOMASZ JANOWSKI

Delft

Conferencia; DG.O'18; 2018

The goal of this paper is to propose a methodology comprising a range of visualization techniques to analyze the interactions between government and citizens on the issues of public concern taking place on Twitter, mainly through the official government or ministry accounts. The methodology addresses: 1) the level of government activity in different countries and sectors; 2) the topics that are addressed through such activities; 3) the resources shared between government and citizens as part of interactions; 4) the intensity of citizen response to government announcements; 5) the sentiment expressed by citizens when providing such responses; and 6) the combinations of such issues. Example combinations include identifying topics that generated the largest Twitter activity by government but received the least interest from citizens, identifying topics that generated the most polarized reactions from citizens, or determining correlation between policy announcements and trust, fear and other negative emotions expressed by citizens. The methodology uses visual analytics to reveal patterns and trends associated with various questions, complemented with sentiment analysis to study government-citizen interactions on Twitter. The methodology is validated by examining Twitter presence in five sectors --- health, social development, education, environment and work, in five Latin American countries with mature e-Participation capabilities --- Argentina, Chile, Colombia, Mexico and Uruguay.