Learn powerful command-line skills to download, process, and transform data, including machine learning pipeline.
We live in a busy world with tight deadlines. As a result, we fall back on what is familiar and easy, favoring GUI interfaces like Anaconda and RStudio. However, taking the time to learn data analysis on the command line is a great long-term investment because it makes us stronger and more productive data people.
In this course, we will take a practical approach to learn simple, powerful, and data-specific command-line skills. Using publicly available Spotify datasets, we will learn how to download, process, clean, and transform data, all via the command line. We will also learn advanced techniques such as command-line based SQL database operations. Finally, we will combine the powers of command line and Python to build a data pipeline for automating a predictive model.
Downloading Data on the Command Line
-In this chapter, we learn how to download data files from web servers via the command line. In the process, we also learn about documentation manuals, option flags, and multi-file processing.
Data Cleaning and Munging on the Command Line
-We continue our data journey from data downloading to data processing. In this chapter, we utilize the command line library csvkit to convert, preview, filter and manipulate files to prepare our data for further analyses.
Database Operations on the Command Line
-In this chapter, we dig deeper into all that csvkit library has to offer. In particular, we focus on database operations we can do on the command line, including table creation, data pull, and various ETL transformation.
Data Pipeline on the Command Line
-In the last chapter, we bridge the connection between command line and other data science languages and learn how they can work together. Using Python as a case study, we learn to execute Python on the command line, to install dependencies using the package manager pip, and to build an entire model pipeline using the command line.