Julia, Python and R compared (draft)

This post contains some comparison notes between these three languages.

DataFrames.jl vs. pandas vs. dplyr

Comparison from DataFrames.jl user guide

Reducing Pandas memory usage #1: lossless compression Reducing Pandas memory usage #2: lossy compression https://pythonspeed.com/articles/data-doesnt-fit-in-memory/ https://pythonspeed.com/datascience/ https://pandas.pydata.org/docs/user_guide/gotchas.html#dataframe-memory-usage https://www.dataquest.io/blog/pandas-big-data/

Performance using packages and running code

• Julia is slower than Python or R when loading packages in a script or during initial running of functions due to precompiling.

• When executing code Julia is faster than Python or R because it is compiled to native code. When Python or R call code implemented in a package that's compiled to native code, the speeds are similar.

An example task is detailed on the following page: Simple moving average speed comparison in Julia, Python and R

Calling code in one language from another

• reticulate - Call Python from R

• rpy2 - Call R from Python

• PyCall - Call Python from Julia

• PyJulia - Call Julia from Python

• RCall - Call R from Julia

• JuliaCall - Call Julia from R

Resources

Julia vs. Python: Which is best for data science? (May 2020) Python vs. R for Data Science: What’s the Difference? (January 2020)