Privacy settings within dsTidyverse

Data owners can manage privacy settings in two main ways: (i) setting the privacy control level, and (ii) controlling which additional functions can be passed to ds.tidyverse.

Privacy control levels

DataSHIELD implements privacy control levels, which allows data owners to control which functions can be used by researchers. The table below shows which dsTidyverse functions are permitted in which privacy mode. This option can be set on the server: for example to set to non-permissive mode use datashield.privacyControlLevel = "non-permissive"

Function Permissive Banana Avocado Non-Permissive
arrangeDS
asTibbleDS
bindColsDS
bindRowsDS
caseWhenDS
distinctDS
filterDS
groupByDS
groupKeysDS
mutateDS
renameDS
selectDS
sliceDS
ungroupDS

Permitted functions

dsTidyverse allows additional functions to be passed via the tidy_expr argument. For example, using ds.mutate you can pass as.numeric:

ds.mutate("mtcars", list(cyl = as.numeric(cyl)), "newobj")

Functions are only allowed to be passed which do not risk disclosing individual level data. The default list of allowed functions is:

"everything", "last_col", "group_cols", "starts_with", "ends_with", "contains",
"matches", "num_range", "all_of", "any_of", "where", "rename", "mutate", "if_else",
"case_when", "mean", "median", "mode", "desc", "last_col", "nth", "where", "num_range",
"exp", "sqrt", "scale", "round", "floor", "ceiling", "abs", "sd", "var",
"sin", "cos", "tan", "asin", "acos", "atan", "c", "as.character", "as.integer",
"lag", "diff", "cumsum"

These defaults can be managed by the data owner on their server using the option tidyverse.permitted.functions. For example, if as a data owner you want to restrict the permitted functions to as.numeric and as.integer, you can set the option tidyverse.permitted.functions = c("as.numeric", "as.integer").

WARNING: This feature gives the data manager the option to restrict allowed functions, but also to allow additional functions. If you choose to allow functions not included in the default list, please take steps to ensure that they are compatible with your research setting, and in sensitive settings with secure data they do not risk returning individual level data to the researcher. If you have doubts please contact the maintainers of dsTidyverse who can discuss the risks with you.