Spatiotemporal Kernel Density Estimation

Spatiotemporal kernel density estimation (STKDE) extends traditional KDE by integrating spatial and temporal kernels to capture dynamic clustering patterns of events. Widely used in crime analysis, STKDE enables predictive mapping of “hotspots” that vary across both space and time. Enhancements such as cyclic STKDE (cSTKDE) further improve accuracy by modeling periodic crime rhythms. Case studies from Baton Rouge demonstrate STKDE's superiority over spatial‐only methods in forecasting burglary and robbery hotspots. Ongoing developments include real‐time, network‐based, and machine‐learning‐integrated STKDE, with applications extending to epidemiology and urban safety.