Infrastructure is a crucial consideration for creating code that can be reused and repurposed. Infrastructures of data are highly rhetorical, because they structure how others interact with information and reflect hierarchies of information, people, and things. Those creating open source and open science projects, in particular, must consider how they structure their data according to principles of accessibility, reusability, and replicability. Failure to consider the design of publicly available code and insufficient user documentation lead to issues with transparency, learnability, and usability within open source science projects.This paper explores the decision-making processes of two open science researchers as they design their data according to the principles of open science. Using interviews and document analysis, I analyze the rhetorical coding choices of researchers and how these choices encourage or discourage accessibility for different audiences. Ultimately, I examine what accessibility means for these researchers working under the umbrella of open science.