A Two-Stage Cox Process Model with Spatial and Nonspatial Covariates

There are rich new marked point process data that allow researchers to study disparate problems such as the factors affecting the location and type of police use of force, and the characteristics that impact the size and location of forest fires. We develop a novel modeling approach for marked point processes that allows for both spatial and nonspatial covariates; both types of covariates are present in the examples we consider. Via simulated and real data examples, we find that our two-stage log Gaussian Cox process model is flexible and easy to interpret, and potentially useful in many areas of research.