Developed a machine learning model (evaluated Decision Trees, Logistic Regression, SVM, Naive Bayes, KNN, Random Forest, Light GBM, XGBoost) to classify shooting incidents as fatal or non-fatal, providing actionable insights for targeted policing strategies. Exploratory Data Analysis included bivariate analysis (Cramer’s V, Chi-square, Mann-Whitney U, KDE plots) and time series analysis with an interactive map. Feature encoding involved ordinal, one-hot, and target encoding; data imbalance was addressed using SMOTE and Random Under Sampler. Support Vector Machine (SVM) was selected as the superior model due to its higher accuracy, recall, and precision in capturing complex relationships.