This review discusses frontier topics in economic geography as they relate to firms and agglomeration economies. We focus on areas where empirical research is scarce but possible. We first outline a conceptual framework for city formation that allows us to contemplate what empiricists might study when using firm-level data to compare the functioning of cities and industries with each other. We then examine a second model of the internal structure of a cluster to examine possibilities with firm-level data for better exposing the internal operations of clusters. An  overwhelming theme of our review is the vast scope for enhancements of our picture of agglomeration with the new data that are emerging.