In my dissertation project, I aim for an integration of two hitherto disjunct research methodologies: the aggregational approach to language variation as exemplified by dialectometry, and the probabilistic modeling of language variation and usage frequencies customarily utilized by, e.g., variationist sociolinguists and probabilistic grammarians. The project will consist of both (i) the development of innovative, hybrid methodologies along these lines and (ii) an empirical assessment of these on the basis of a corpus-based application to geographic variability in English morphosyntax. The proposed approach is, however, intended to be generally applicable to all those cases where large-scale patterns in multi-feature frequency-based datasets take center stage, such as in historical linguistics and studies of register variation.