Analysing User Preferences and Trends on Social Media through Big Data and Machine Learning Techniques

Abstract

Social media platforms have become integral parts of everyday life for billions of users worldwide. The vast amount of user-generated data from these platforms offers valuable insights into user preferences, opinions, trends, and online human behavior. This paper provides a comprehensive review of how big data analytics and machine learning can be leveraged to understand social media users and content. We first highlight popular social media platforms and the vast, diverse data they generate. We then discuss data collection, pre-processing, storage, and analysis methods tailored for the scale and noise of social data. Next, we review applications of big data and machine learning, from predicting trends and events to personalized recommendations. We also examine various data mining techniques, statistical modeling approaches, neural networks, and other algorithms commonly used. Ethical considerations around privacy, transparency, and algorithmic bias are also addressed. Three case studies showcase social data mining in practice for trend analysis, user profiling, and visualization. Finally, current challenges and future directions are discussed around scalability, noise reduction, explainability, and stream analysis. Overall, this paper serves as a guide for academics, analysts, and platforms interested in maximizing insights from user data while minimizing harms through thoughtful modeling techniques and responsible practices.