BIG DATA: APLICACIÓN AL ANÁLISIS DE SENTIMIENTOS EN AEROLÍNEAS

Abstract

Discovering new forms of data collection has been possible thanks to the rise of social networks and technology 2.0. In this research was carried out a bibliographic review on the concept of Big Data and the techniques used for its analysis. Moreover, was analyzed and predicted consumer’s sentiments in airlines using the Knime data mining platform. To do this, a model based on decision trees has been chosen, checking their reliability through confusion matrices. The used data comes from the Quandl database in which, through algorithms, more than 6000 websites are scanned every 5 minutes. The data set includes the period from January 2013 to June 2017. The obtained results show scoring intervals that reflect the weighted average sentiment of the consumer with respect to the initial sentiment and the volume of existing news. In this line, we observe how polarized high and low feelings carry extreme intraday weighted average feelings. It should be noted that the model used in the study predicts feelings with 91% reliability in the case of the airline Ryanair and 77% in SouthWest Airlines.