Real-Time Analysis of Women’s Safety in Indian Cities Using Twitter Data and Machine Learning

Abstract

This paper addresses the continuing problem of ladies’s protection in metropolis India, wherein harassment and assault incidents are regularly underreported due to systemic and procedural constraints. It proposes a unique framework that leverages Twitter as an actual-time, crowdsourced sensor for taking photographs public sentiment and protection perceptions. Using herbal Language Processing (NLP), device studying (ML), and Geospatial assessment, the device classifies sentiment, identifies ordinary topics, and visualizes spatial forms of protection worries. Tweets are gathered through Twitter APIs, preprocessed, and analyzed using classifiers which consist of Naive Bayes and useful resource Vector Machines. Problem depend modeling via Latent Dirichlet Allocation (LDA) well-known shows dominant worries. This paper contributes to information-pushed metropolis protection planning via allowing actual-time, citizen-powered public protection tracking.