In this article, we propose a method for conducting likelihood-based inference for a class of nonstationary spatiotemporal log-Gaussian Cox processes. The method uses convolution-based models to capture spatiotemporal correlation structure, is computationally feasible even for large datasets,...