Latent Profile Analysis of Crime Patterns: Insights from Spatial, Temporal, and Behavioral Data

This study employed Latent Profile Analysis (LPA) to uncover distinct crime patterns using spatial, temporal, and behavioral data from Pampanga, Philippines, spanning 2017 to 2020. A Gaussian Mixture Model identified four optimal profiles based on the Bayesian Information Criterion (BIC). This approach revealed trends in seasonal, temporal, and contextual crime dynamics. Results indicated that Profile 1 reflects pandemic-driven shifts in 2020 with reduced crime during nonholidays and the Dry season. Profiles 2 and 4 highlighted cyclical crime patterns linked to economic activities such as holidays and paydays. Profile 3 showed high-frequency crimes in 2019 tied to the government efforts related to illegal drug cases. Spatial analysis further identified hotspots in urban centers in the Pampanga. This study demonstrated the ability of LPA in revealing deeper insights into crime trends by identifying hidden patterns and relationships across diverse variables.