Women Safety Platform with Safe Route Prediction Using Crime Data

Women's safety in public spaces is a pressing concern, particularly in urban environments where traditional navigation systems like Google Maps fail to prioritize safety over efficiency. This paper proposes a Women Safety Platform that leverages geospatial data, machine learning models, and user feedback to address these limitations. The platform predicts and highlights the safest route between source and destination by analyzing historical crime data and assigning safety scores to streets using machine learning models, including K-Nearest Neighbors (KNN), Random Forest, Logistic Regression, Decision Tree, and Bernoulli Naive Bayes. KNN effectively identifies patterns in crime data to provide accurate safety scores, while Random Forest and Decision Tree classifiers achieved perfect training accuracy, showcasing their ability to capture complex patterns, albeit with potential risks of overfitting. Logistic Regression demonstrated a balance of high accuracy and generalization, and Bernoulli Naive Bayes provided reliable performance with the least overfitting tendency. Routes are evaluated based on aggregated street safety scores, visually marked to aid user decision-making and complemented with sentiment-based insights from user feedback. The Women Safety Platform uses geolocation data to monitor user safety and triggers an emergency SOS if safety levels drop below a threshold. By integrating crime data, real-time geospatial APIs, and proactive emergency measures the Women Safety Platform empowers women with a navigation tool that prioritizes safety alongside convenience. This initiative not only fosters confidence and freedom in navigating public spaces but also serves as a scalable model for enhancing mobility and safety for other vulnerable populations.