The increasing possibilities for the study of specialized discourse have seen terminologists dealing with larger and more heterogenous corpus data, a context in which ready-made tools might fall short. This work investigates if the R programming language can help researchers analyze complex specialized corpora and uncover clues of diastratic variation in language for specific purposes, a phenomenon understood as the coexistence of different language uses among groups of experts in the same field. Since one selling point of R is its capacity to stay up to date with techniques in both statistics and machine learning, an experiment is proposed in each of these areas: one dives into the exploratory analysis of categorical data, while the other relies on distributional semantics and deep-learning technology. Together, these series of tests make it possible to discuss key perspectives and limitations for R in terminology studies.