sparklyr

Allows using familiar dplyr verbs like filter and group_by directly on Spark DataFrames, enabling R users to scale tidyverse workflows without learning new syntax, as shown in the flights filtering example.

Provides interfaces to Spark ML algorithms for tasks like linear regression, with support for feature transformations and tuning, demonstrated in the ml_linear_regression workflow with mtcars data.

Enables running arbitrary R code across clusters using spark_apply(), allowing custom analyses and package usage in a distributed manner, as illustrated with the broom::tidy example for linear models by species.

Offers built-in tools in RStudio for managing Spark connections, browsing tables, and previewing data, enhancing interactive development with a dedicated Spark pane and connection dialogs.

Supports building custom extensions to call full Spark APIs or integrate third-party packages, increasing flexibility beyond core functionality, as shown in the line counting example.