Evaluación de la percepción de los servicios públicos en Colombia mediante Text Mining vía Twitter

Abstract

In Bogota, the public service providers are Enel, Vanti and Acueducto, big companies that are characterized by a customer centricity operation. However, these companies are the ones with the highest number of complaints and claims reported to regulatory entities. The perception of service may not be the best. Determining the veracity of this fact is the objective of this research, through the exploitation of text mining techniques by taking advantage of the voice of the customer in the user’s tweets; applying the Knowledge Discovery in Databases methodology to generate the database composed of 9071 tweets of these three companies. In data cleaning phase, additional steps are established to refine the database and consolidate the tweets of interest for the research. This allows to obtain an exploitation of the data explaining the results through word clouds, frequency diagrams, sentiment analysis and the ratios between the polarity of the tweets. The results enable inferring that, for the time interval in which the analysis was performed, the perception of the service is not good and so, there are opportunities for improvement for the three companies.