Facebook Text Mining and Sentiment Analysis using R: a case study of two Italian companies

Abstract

Today, everyone knows and uses social media. Social media offers a powerful environment for people’s thoughts and feelings and it is an enormous ever-growing source of data ranging from everyday observations to involved discussions. Moreover, it is possible for marketing researchers to analyze these data using two techniques: Text Mining and Sentiment Analysis. The aim of the research reported in this thesis is to analyze the reaction of the Italian population on Facebook about the launch of two new products in the market and, on the other hand, to evaluate the communication solution strategies of two Italians brands. An important step for a company during the launch of a product is the online advertising campaign. A promotional campaign on Facebook is useful to obtain information from user behavior and to understand what they publish on the web. It is necessary to point out that the launch of a new product can lead the company to have an increase in its visibility and consequently an increase in its profit. Therefore, the marketing and advertising action taken during the period preceding the day of launch becomes important. In today's connected world, we have witnessed a strong technological evolution that has led the Web to become a new communication channel. Thanks to the use of digital technologies, people create their own content, retrieve information they want and send messages at any time. Social media constitute a subject challenging source of new information gathering and opinion making. Big companies exploit the potential of social media during the product launch phase in order to directly reach the target audience and subsequently be able to analyze their contents. Try to study the needs, opinions and behaviors of users through the analysis of Social Media has revealed the need to develop new methodologies for the collection, modeling and analysis of data extracted online. These methodologies are: Text Mining and Sentiment Analysis. The tools of Text Mining and Sentiment Analysis are useful to the company to better understand if the actions carried out online have been appreciated by the user. The purpose of Sentiment Analysis, is to discover the "feelings" or the opinions and attitude of the user on a theme through the analysis of the texts published and shared by the users. Text mining, on the other hand, is useful for analyzing and extrapolating textual information on the web and on Social Networks. Both disciplines have found a strong application in the field of marketing and brand management. In this thesis has been analyzed the potential of the two methodologies to study two Italian business cases. The theoretical analysis will start from the concept of development on the use of Social Media and will be applied to the world of the food industry which will be the one related to the two brands analyzed in the case study. A focus will be left to Facebook because it will be the social media used to collect data. In the second chapter will be analyzed the world of data mining, introducing Sentiment Analysis and Text Mining. The study will start from an introduction and then will be described the processes, methodologies and the software used “R”. The third chapter will be devoted to the description of the characteristics of the Social Network of reference, Facebook, and how the data extracted from it can be used for Text Mining and Sentiment Analysis. The last chapter will be the most practical.The collected data will be analyzed, through a dataset. The analysis time period will be related to two distinct moments regarding the launch period of the two products: Nutella Biscuits and Crema spalmabile Pan di stelle. Subsequently the analysis will be made on the two methodologies previously described. The thesis will conclude with a comparison between the results obtained and with personal reflections.