Reference guide: Python for machine learning

Previously, you learned more about the Python ecosystem for machine learning. You developed
an understanding of the different Python file types that are available to approach a data
analytical task, along with the various types of integrated development environments in which
the coding takes place.

Here, you will learn more about different Python use cases, along with the advantages and
disadvantages of using different types of files and development environments.

Coding languages for data professionals

As the field of data science has progressed, steadily tools have been released to facilitate the
development of data-driven solutions for various problems. Not only have the tools become
more sophisticated, but often easier to use as well.

Pieces of software such as Tableau and Looker have made performing data analysis much
more simple and efficient. They can allow for quick yet comprehensive overviews of a dataset
and are often used as a starting point before deeper analyses or developing models from the
data.

This is where coding languages have a huge role in continuing to solve the task at hand. Not
only can they perform much of the same preliminary analysis that is done in other pieces of
software, but they also contain some very powerful functionality that can be used quite easily.

The two most popular coding languages for data science are R and Python. For the most part,
both achieve all the same things, and deciding which one to use usually depends on personal
preference or what the rest of your team is using.

The R language was designed for statistics from its inception. Much of the functionality required
is baked into the language itself. Python, on the other hand, was and still is a general purpose
language that gained popularity within the data science ecosystem.
In this certification, you have exclusively used the Python language to learn about data analytic
techniques and how to use those techniques to solve problems you might encounter in the
workplace. While you might come across a situation where you need to use R to perform an
analysis, the same principles and concepts you have learned here apply no matter the software
used.

Types of Python files

Python scripts

Python scripts are arguably the more common type of file overall, but not necessarily in the
world of data science. They are denoted with the file extension “.py,” and are used much more
for larger projects, or for projects where it is not essential to see each part of code run
separately. For code that is going to be deployed and put into production, Python scripts are
much more common. They are much easier to debug than Notebooks, along with being much
better for reproducing results. On top of that, they work much better with other pieces of
software and infrastructure.

Python notebooks

Python notebooks are what you’ve been using throughout this entire program. While they
technically can do everything that a script can do, they are mainly used for exploration,
visualization, or presentation.

One of the main features of Python notebooks is being able to easily run different sections of
code independently. Additionally, you are able to see the output of each section of code, rather
than having it all come out at the end as is the case with scripts.

Notebooks also contain functionality to insert non-code elements into the notebook itself. If you
want to insert text inside a Python script, you are limited to writing comments. But with
notebooks, you are able to add markdown text, images, and links to provide more context to the
code.

Integrated development environments for Python

When doing any coding related task, an integrated development environment, or IDE, is where
much of the work is actually done. It is a piece of software that gives a place to write, test, and
run code. For any programming language, there are often many different IDEs that are available
to use, varying slightly in functionality and included tools. Selecting one to use often comes
down to what types of tools you need to create your program or even just personal preference.

In this certificate program so far, you’ve been using Jupyter Notebooks on the internet. Jupyter
Notebooks is an IDE, however it only supports Python notebook files. It is possible to create
your own instance of Jupyter Notebooks on your personal device, but the functionality is
essentially the same as using a web-based instance. Other IDEs, such as Visual Studio Code
and PyCharm, run locally on your device.

IDEs can have many included tools, but there are a few that are so common that you’ll find them
paired with almost every IDE. Code completion, file management, and debugging support come
in very useful, streamlining your workflow and letting you solve problems in your code as they
come up.

Key takeaways

● Coding languages are useful for approaching a data-driven problem. Python and R are
two popular options for coding languages, each with their own advantages.

● The two main types of Python files are known as Python scripts and Python notebooks
○ Python scripts are better for production grade code, and are easier to debug and
manage.
○ Python notebooks are better for exploratory analyses, presentations, or anything
that needs to be human-facing. You are able to insert images, text, and links
directly into the code.

● An integrated development environment (IDE) is a piece of software that gives a place to