dsTidyverse: An implementation of Tidyverse within the DataSHIELD ecosystem

• DOI: 10.31219/osf.io/hwpdz_v1
• Publication Date: 2025-02-06T19:19:24Z
• Authors: Tim Cadman; Mariska Slofstra; Demetris Avraam; Eleanor R A Hyde; Niels Kikkert; Marije van der Geest; Dick Postma; Ruben Veenstra; Stuart Wheater; Erik Zwart; Morris Swertz

This paper introduces dsTidyverse, an R package designed to enhance data handling within the federated analysis platform DataSHIELD. DataSHIELD enables multi-site analysis without direct data sharing, crucial for privacy-sensitive research. While DataSHIELD facilitates complex analysis, it lacks user-friendly data manipulation tools. dsTidyverse addresses this by implementing selected functions from the “Tidyverse” ecosystem within DataSHIELD’s client-server architecture. The package provides functionality for selecting, renaming, and creating columns; conditional recoding; combining data frames; filtering rows; grouping data; and converting to tibbles. Rigorous disclosure checks are implemented to prevent individual-level data leakage. The paper demonstrates, through examples, how dsTidyverse simplifies common data manipulation tasks, improving user experience and analysis efficiency within DataSHIELD. The package is open-source, freely available on CRAN and GitHub, welcoming further development. See https://github.com/molgenis/ds-tidyverse