when-to-use-num-vs-other-tools.md

When to use Num vs. other tools

Use Num when you have small data, or small tasks, or one-time needs, or development needs. For example, if you want a quick one-time result on your command line, then Num is good.

Use other tools when you have large data, or large tasks, or oft-repeating needs, or production stability needs.

When to use datamash, or qstats, etc.

We suggest these small compiled binaries when:

• You're a sysop or sysadmin.
• Your needs include systems automation.
• You think of numbers in terms of streams and pipes.
• You already use system tools, such as awk, grep, sed, make, etc.
• Your data set fits in available RAM and is suitable for Unix pipes.
• You want a quick way to pipe information to commands, without any dependencies.
• Your ideal tools are small, compiled once, with no dependencies, and no add ons.

When to use R, or Julia, etc.

We suggest R and its tools, such as CRAN and R Studio, when:

• You're a statistician.
• Your needs include doing data exploration.
• You think of numbers in terms of vectors and functions, such as matrix mathematics.
• You already use a statistics tool, such as R Studio, Matlab, or Mathematica.
• Your data set fits comfortably in your computer's memory.
• You want a quick easy way to try visualizations and algorithms for yourself.
• Your ideal tools are on the leading edge of new statistics.

When to use Python, or Scala, etc.

We suggest Python and its tools, such as pandas and scikit, when:

• You're a coder.
• Your needs include doing data pre-processing or post-processing.
• You think of numbers in terms of objects and messages, such as OOP methods.
• You already code in Python, Scala, Java, Perl, Ruby, Go, etc.
• Your data set exceeds your computer's memory.
• You want production environment deployments of visualizations and algorithms.
• Your ideal tools use a general purpose programming language.