Social Media Survey Dataset: Exploring the Impact of social media on Personal Well-Being

Abstract

Social media has become an integral part of life, and step by step occupies a certain position in people’s everyday lives. Concerning its recourse in social life, it has created rising societal concern in areas such as general happiness, mood, concentration, and sleep. The research, therefore, systematically and comprehensively explores these effects with the help of the Social Media Survey Dataset, which consists of cross-sectional ordinal data of various characteristics and potential behavioral outcomes of employing such platforms. This research utilizes data collected from participants of different ages and various jobs. Descriptive questions concerned with specific content, the feelings evoked from content, and likely impact of content consumption on mood, attention span and quality of sleep. Thus, the data was analyzed using such steps of data preprocessing as feature encoding and dimensionality reduction using PCA. For the initial analysis of the project, five models of machine learning were employed which include Logistic Regression, Support Vector Machines, Random Forest, XG Boost, and Artificial Neural Network. The first significant set of performance measures was used to find a reference point for the following studies and to determine to what degree modern models can identify patterns in the explored dataset. Classifiers depicted a moderate baseline accuracy between 25% to 59% suggesting that there could be an improvement. These results shed the light on the dataset for future research on the impact of social media and to enhance the finding of the intervention in the field of digital health and wellness.