NO. 2(2026) 228 Received: 07-04-2026 | Accepted: 15-05-2026 | Published: 22-05-2026 | www.ajaccm.com Unravelling Social Media Patterns Through Data Insights

Abstract

Massive amounts of user-generated data are produced by social media platforms, reflecting the opinions, feelings, and behavioral patterns of the general population. Because this data is unstructured and changing, it is difficult to extract significant insights from it. The goal of this project, "Unravelling Social Media Patterns Through Data Insights," is to use cutting-edge data-driven methods to analyze and interpret social media data from sites like Facebook, Instagram, and Twitter. The system classifies user sentiments and groups conversations into relevant subjects by utilizing Natural Language Processing (NLP) and machine learning models, such as TF-IDF and BERT embeddings. The performance of several classifiers, including Random Forest, Logistic Regression, and Deep Learning models, is assessed. Key Words: Social Media Analytics, UserGenerated Data, Public Opinion Analysis, Sentiment Analysis, Natural Language Processing (NLP), Machine Learning, TF-IDF, BERT Embeddings