Abstract
The COVID-19 context affected the use of social media. Video and voice chat facilitate social interactions during the current social distancing requirements. However, social media creates unrealistic reference points of comparison. The time spent on social media can thus diminish well-being. Researchers and managers aspire to understand how sentiments can control social media. Another research interest regards which techniques create positive sentiments and enhance user experience. This chapter introduces the main stakes of social media, how sentiments change social media, and in turn, social media influences sentiments. The main focus presents a literature review regarding the techniques to analyze sentiments. Finally, solutions and recommendations contemplate the use of social media, for both users and social media platforms.
TopIntroduction
Natural disasters motivate researchers to analyze users’ behavior and sentiments on social media (Gao et al., 2020; Pathak et al., 2020). The Covid-19 pandemic has changed the way people used to live and behave on social media (Albahli et al., 2020). The situation heightened mental and physical issues due to diseases and stress (Campbell & Gavet, 2021). Social media gives various ways to communicate and create social bonds. It represents a timely concern to fight against isolation. People spend more time on social media, willing to enhance their well-being and social life (Boyd & Ellison, 2007; Nyagah et al., 2015). However, users’ perceptions can be wrong. Research has shown that the use of social media enhances signs of depression, anxiety, and sleeping disorders (Milyavskaya et al., 2018; Utz et al., 2015; Verduyn et al., 2015). People tend to do upward comparisons, diminishing their self-esteem whereas most people on social media post filtered pictures (Hamasaki et al., 2009; Muqaddas et al., 2017). In addition, the development of video and voice chat facilitates social interactions, making them more realistic and human. More than text, the voice, and facial expressions enhance sentiments (Dai et al., 2015). Therefore, social media can analyze users’ attitudes to improve user experience, and in turn sentiments and behaviors (Albahli et al., 2020). In marketing, researchers and managers can conduct tests to understand consumers’ behavior and their level of trust in a brand's message (De Keyzer et al., 2017). Neuroscience and artificial intelligence techniques can analyze users’ sentiments and behaviors during their social media experience (Zhang et al., 2020).
This chapter aims at explaining (1) the stakes of social media, (2) the way sentiments influence social media and in turn, social media influences sentiments, and (3) techniques to analyze user’s sentiments. The first part presents the background of this chapter, with social media characteristics and the stakes of social media; the second part focuses on social media and sentiments, the type of data necessary to do sentiment analysis, and the techniques of sentiment analysis on social media; the third part suggests solutions and recommendations regarding unhealthy social comparisons and risky behaviors on social media, as well as solutions for social media platforms to develop a user-centric strategy; finally, the fourth part brings out future research directions regarding new ways of conducting sentiment analysis on social media, like media ethnography or neuromarketing, and discusses the role of social media moderators.
Key Terms in this Chapter
Ethnography: A research method used by sociologists to study and comprehend groups, organizations, and communities.
Prosody: The melody of the voice represented by the acoustic and non-verbal parameters of voiced speech usually defined by the fundamental frequency, rhythm and intensity of the speech signal.
Natural Language Processing: Opinion mining technique with extracting information about people’s thoughts and feelings from a corpus of text data (Albahli et al., 2020 AU97: The in-text citation "Albahli et al., 2020" is not in the reference list. Please correct the citation, add the reference to the list, or delete the citation. ).
Sentiment Analysis: The field of study that analyses people’s opinions, sentiments, evaluations, appraisals, attitudes, and emotions towards products, services, organizations, individuals, issues, events, topics, and their attributes (Almuraih et al., 2020 AU98: The in-text citation "Almuraih et al., 2020" is not in the reference list. Please correct the citation, add the reference to the list, or delete the citation. ).
Social media: Online networks through which users can create online communities to share information, ideas, personal messages, and other content like videos.
Sentiment: An emotional state occurring as the result of an emotion through external or internal causes (i.e., happy and joyful, or painful and sad).
Data Mining: Family of tools allowing the analysis of a large amount of data on social media.
Subjective Lexicon: A word list nominated to a score that shows its nature in terms of positive, negative or objective opinion ( Alessia et al., 2015 ).
Frequency: Number of oscillations per second expressed in Hertz (Hz). For the voice, the frequency corresponds to the number of opening and closing phases per second of the vocal folds.