URBANGUARD AI: PROTECTING WOMEN THROUGH SOCIAL MEDIA SENTIMENT INSIGHTS

Abstract

Ensuring the safety of women in urban areas remains a critical societal challenge, and real-time insights are essential for effective policy and intervention strategies. Social media platforms, particularly Twitter, offer a valuable source of public opinion and firsthand experiences that reflect safety concerns in cities. This study introduces WomenSafe AI, a machine learning– based framework for analyzing Twitter sentiments to investigate women’s safety in Indian cities. The system leverages natural language processing (NLP) for data preprocessing, feature extraction, and sentiment classification, combined with supervised learning models to identify patterns of concern and areas of heightened risk. Experimental results demonstrate that WomenSafe AI can accurately categorize sentiments as positive, neutral, or negative, providing actionable insights into public perceptions of safety. The findings highlight the potential of social media–driven analytics to assist policymakers, law enforcement, and community organizations in understanding urban safety dynamics and implementing targeted interventions.