• FeatureTransforms.jl has transforms, but they are not fully revertible. Some of their transforms such as MeanStdScaling are constructed for a specific table and cannot be inserted in the middle of a pipeline for example.
  • AutoMLPipeline.jl relies on the Python stack via PyCall.jl. They provide pipelines with Julia's pipe |> operator and follow a more "Pythonic" interface. They do not support general Tables.jl.
  • Impute.jl, Cleaner.jl, DataConvenience.jl all have a small set of transforms related to fixing column names as well as other basic transforms that we plan to absorb in the long term.
  • DataFramesMeta.jl is a package to manipulate DataFrames.jl tables. It is not intended for statistical transforms such as PCA, Quantile, etc, which rely on complex interactions between the rows and columns of a table. The usage of macros in the package promotes one-shot scripts as opposed to general pipelines that can be passed around to different places in the program.
  • Query.jl is a package to query IterableTables.jl. Similar to other alternatives above, the package is not intended for advanced statistical transforms.
  • MLJ.jl is a popular machine learning framework in Julia that adopts a different design for pipelines based on mutability of structs.