Minería de datos y análisis de sentimientos en redes sociales, caso de estudio: percepción del covid-19 en el Ecuador.

Abstract

Sentiment Analysis, SA, Sentiment Analysis is of great importance in determining people's attitudes, emotions, opinions and comments. There are several cases of application of SA; For example, in the case of companies, it can be useful to determine the preferences of their customers for a certain product or service, In the government sector SA can be used to study the satisfaction of citizens with respect to its management, Another case of application of SA is in the entertainment and entertainment industry such as the cinema, and it can be used to analyze the opinions of the public. Sentiment analysis is commonly done through comments from social networks such as, Facebook, Instagram, Twitter, YouTube, etc. The purpose of this work consisted in the analysis of comments retrieved from the social network Twitter in the context of COVID-19 in Ecuador, from April 2020 to July 2021. The methodology applied to develop this project was based on the best practices of Cross Industry Standard Process for Data Mining, CRISP-DM, from IBM and Team Data Science Process, TDSP from Microsoft, The phases and activities developed are, 1 Business understanding, which consists of establishing tools and repositories of comments to be obtained from Twitter. 2 Data acquisition and understanding, which refers to the pre-processing and cleaning of data. 4 Modeling involves the exploration, processing, analysis and visualization of data. 5 The evaluation consists of the comparison of models through the calculation and analysis of the metrics. And finally, 6 The deployment that tries to make the decision to choose the most appropriate method of sentiment analysis. The first results obtained were for the most frequent words, It was analyzed by time segments, 1 In the period between April and July 2020, the most frequent words were, COVID and fear, implying that the population was entering a stage of panic. 2 Then in the period between August and November 2020, the words, Pandemic, Cases, Funeral homes, overflowing, rotten stood out, which denotes that very difficult moments of losses were being experienced human rights and a strong uncertainty in the country. 3 Subsequently, the period analyzed between December 2020 and March 2021, the most relevant words were, Ecuador, Sum, Cases, when comparing with the statistics in that period, COVID cases actually rose -19 in the country. 4) Finally, in the period from April to July 2021, the most frequently mentioned words were, vaccine, COVID, which coincides with the vaccination process that was carried out in this period.