Subjective well-being and social media

Iacus, S. M.; Porro, G.

This book presents an overview of the most recent projects on the estimation of subjective well-being through social media data. In particular, it focuses on a new project, aimed at constructing a Twitter Subjective Well-Being Index, which started in 2012-almost at the same time of expansion of sentiment analysis to Twitter data-and grew slowly till the present days. The project was originally conceived at the University of Milan (Italy) and then embraced later in 2015 by the University of Insubria (Como, Italy), the University of Tokyo and the University of Waseda in Japan. The book reviews the different approaches to the estimation of well-being, from traditional macro-economic definition-both one-dimensional and multidimensional-to survey analysis and finally to big data and social networking sites (SNS) in particular. It introduces briefly the most commonly used machine learning and statistical techniques for textual analysis. It also serves two scopes: to explain how machines transforms text into meaningful statistics, and also to convey the idea that human supervision is an essential step of this process whatever technique is used. The book presents different SNS-based subjective well-being indexes that have been proposed in the literature, with a special focus on the one proposed by the authors. Among all positive aspects of SNS data, there are also some pitfalls which are quite easy to imagine, and well known to the experts in the field. The main one is that social media accounts/users/data cannot be considered statistically representative of the demographic population. The book presents a possible approach to tackle the selection bias problem by anchoring social media indexes to official statistics. It focuses on the analysis of the impact of the COVID-19 pandemic, that hit the world in 2020, on the social media indexes of subjective well-being. (PsycInfo Database Record (c) 2021 APA, all rights reserved)