Abstract

he vertiginous development of technology and knowledge globalization has generated a high interest on social networks within organizations, where its presence has multiplied exponentially in recent years. That is why, it is proposed in this paper to analyze data extracted from social networks, specifically from Twitter, aiming to obtain different data elements that allow management and analysis of the opinions provided by users on different. This information is very useful for client management and acknowledging preferences of brands and organizations. In particular, this work responds to the following research questions: (1) How does the Twitter user behave in different brands? and (2) How do opinions on the network affect the company’s Twitter profile? In this manuscript, we present the tweet user profile for information analysis via a practical software architecture proposal, which is composed by four layers (extraction of the data source, ETL processes (extraction, transformation, and loading), selection of database, and visualization of the results). The implementation and dashboards of this architecture come from the study case of different types of organizations: banking, telephony, shopping, and supermarkets. The processing of the data corresponds to the extraction of tweets generated by the Twitter users of the organizations. Then, the ETL process is obtained via the useful Spoon from Pentaho Data Integration. The processed data is employed to build the final database, and finally, the generated information is visualized by utilizing dashboards from Qlik Sense Desktop. The results of this study evidence that it is possible to implement a practical architecture to analyze the model information of the Twitter user profile through dashboard; consequently, the organizations can opportunely realize better decisions.