Sentiment analysis in Tripadvisor: the consumer's voice as a decisive parameter in measuring the quality of information in museums

Abstract

The aim of this article is to identify parameters that assist museum managers in measuring the quality of information of their institutions. For this, 1,900 portuguese-language reviews were recovered from September 2017 to October 2019 posted on the TripAdvisor platform, referring to the Museu do Amanhã, located in Rio de Janeiro - RJ.  From a content analysis, reviews not related to the theme "museal information" were excluded, leaving 701 comments. Through the iFeel web system, 18 different methods of sentiment analysis were rotated and assertiveness tests were applied in each of them. The Umigon method was considered the most appropriate in the categorization of positive, negative and neutral opinions for the database worked. The greater number of negative feelings about the quality of information in the Museu do Amanhã relates to misinformed employees, disorganized institutional information and the absence of internal signaling. Regarding positive feelings, it was noticed that the relevance of information transmitted by the main exposure and the high levels of interactivity for retrieving information from the collection were the main favorors. The results also allowed the construction of parameters that can guide decision makers of several other museums regarding the theme quality of information.