Book Image

Hands-On Data Science with the Command Line

By : Jason Morris, Chris McCubbin, Raymond Page
Book Image

Hands-On Data Science with the Command Line

By: Jason Morris, Chris McCubbin, Raymond Page

Overview of this book

The Command Line has been in existence on UNIX-based OSes in the form of Bash shell for over 3 decades. However, very little is known to developers as to how command-line tools can be OSEMN (pronounced as awesome and standing for Obtaining, Scrubbing, Exploring, Modeling, and iNterpreting data) for carrying out simple-to-advanced data science tasks at speed. This book will start with the requisite concepts and installation steps for carrying out data science tasks using the command line. You will learn to create a data pipeline to solve the problem of working with small-to medium-sized files on a single machine. You will understand the power of the command line, learn how to edit files using a text-based and an. You will not only learn how to automate jobs and scripts, but also learn how to visualize data using the command line. By the end of this book, you will learn how to speed up the process and perform automated tasks using command-line tools.
Table of Contents (8 chapters)

From dumb Terminal to glam Terminal

We're going to use gnuplot to render dumb text graphics and canvas-style plots inside our Terminal. To begin, we need some basic configuration for our gnuplot startup. Put the following in ~/.gnuplot:

set term dumb

Next, we need a wrapper around gnuplot to get some fancy graphical output. This wrapper looks at the GNUTERM environment variable of your current shell and does some calculations on the Terminal's width and height so that gnuplot knows how big a window it has. The wrapper will update our ~/.gnuplot configuration with the graphics capabilities specified for our Terminal. We aren't going to delve into the wrapper, but just use it as another command. Here it is:

__gnuplot() {
SIZE=$(stty size 2>/dev/null)
SIZE=${SIZE:-$(tput lines) $(tput cols)}