Python and R for the Modern Data Scientist

Abstract

Computation in many fields including those that use statistical software is increasingly driven by needs that can be addressed in many programming ecosystems. In projects that require statistical analyses, both R and Python comprise two frequent resources. In ecology, R is the most frequently used (Lai, Lortie, Muenchen, Yang, and Ma 2019). In bioinformatic gene set analyses, R is also more frequently used in peer-reviewed publications, but Python is still an important statistical resource depending on the specific project (Xie, Jauhari, and Mora 2021). Python outcompetes other languages in use for machine learning and some forms of factor analyses (Hao and Ho 2019; Persson and Khojasteh 2021; Raschka, Patterson, and Nolet 2020). However, the relative frequency that a tool is used for statistical analyses is only one metric of importance and not necessarily a proxy for its merit or its capacity to support innovation and efficient in analyses for practitioners (Zhao, Yan, and Li 2018). It is thus critical that we explore contrasts of at least these two common software languages that support statistics because data scientists can become isolated or polarized within their specific competencies, ideologies, and workflows. A high-level discussion of strengths and weaknesses specific to data endeavors with statistics is germane to both decisions on specific projects and on competency development as a scientist.