The purpose of this research is to focus on aggressive communication, that gets triggered by emotionally or politically charged events especially in online space. The study investigates whether machine learning models can provide a reliable method for identifying various forms of aggression in Indian Twitter posts, and whether notable patterns of aggressive behaviour are linked to certain categories of events.

In order to understand the above phenomenon, five high-impact events from different domains social, financial, sporting, and political were selected for study and analysis to see their capacity to invoke strong public reactions on twitter.

About 13,000 tweet data related to each of these events was collected using the Python-based snscrape tool, were collected and processed. The aggression was divided in to three categories namely Overtly Aggressive (OAG), Covertly Aggressive (CAG), and Non-Aggressive (NAG)

Out of the four selected supervised leaning models (Random Forest, Support Vector Classifier (SVC), Logistic Regression, and Multinomial Naïve Bayes), Multinomial Naïve Bayes demonstrated the most balanced and effective results, particularly in handling the nuances of covert versus overt aggression. The study showed events related to OAG generated highest volume

This study demonstrates how machine learning can be leveraged to detect and interpret public aggression in complex, multilingual environments like India's digital landscape. Beyond classification, the research provides insight into how aggression manifests across different societal issues and over time. The findings have practical implications for improving online moderation systems, guiding responsible communication policies by the government, and informing future research into the psychology and sociology of digital interactions. By focusing on local context and diverse event categories, the study makes a contribution to computational social science and paves the way for more culturally attuned AI applications...