2. Basic Python

This site provides a number of links to recommended introduction to Python tutorials (no reason to re-invent the wheel by trying to write a new one!). These different tutorials each have their own strengths and weaknesses, and readers are encouraged to shop around a little.

Words of Advice

Before that list, however, some guidance:

Manage your expectations

Python is very user friendly, but it does ask that you invest a little in learning the fundamentals of the language before you jump into applications. Languages like R and Stata were explicitly designed to have you up and running almost immediately, but accomplishing this affected how the languages could be written. Python requires a little training, but pays you back for that investment down the road.

Prioritize

Because Python is a general purpose language, it has many features that definitely won’t be relevant for you immediately, and may never be relevant. Many tutorials, for example, were written by people in silicon valley who write apps for phones, and emphasize the skills they use. Here’s an overview of the things I think are critical to know immediately, important to learn soon, and things you may never need to know:

Need immediately:

  • Data types: integers, floats, strings, booleans, lists, dictionaries, and sets (tuples are kinda optional)
  • Defining functions
  • Writing loops
  • Understanding mutable versus immutable data types
  • Methods for manipulating strings
  • Importing third party modules
  • Reading and interpreting errors

Things you’ll want to know at some point, but not necessary immediately:

  • Advanced debugging utilities (like pdb)
  • File input / output (most libraries you’ll use have tools to simplify this for you)

Don’t need:

  • Defining or writing classes
  • Understanding Exceptions

If you try one of these and find my annotations inaccurate, please send me a note letting me know and I’ll update my advice!

One last note: in the Python world, the type of work you will probably be doing is called “scientific computing”, so if you’re trying to find tailored resources, that’s a great search term!

Interacting with Python

There are three ways you may see people use Python: interactively with Python, interactively with iPython, and by executing a saved file. They’re all basically equivalent, but I want to highlight them so you aren’t surprised if you see them.

Interactive Use: Most social scientists work with Python interactively – you open a session, and execute your code one line at a time, just like in Stata or R.

At its core, Python has a simple, text-based interactive interface. However, there is also a program called iPython many people use when working interactively with Python. iPython is just a little program that sits between the user and Python itself that adds a few bells and whistles to make it a little easier to work with. What iPython can do is not really important here; what is important is that you not be surprised if you find different tutorials working with slightly different interfaces. (If you see a tutorial with something called Jupyter notebooks, that’s basically a special document that runs iPython). You can learn more about iPython and Jupyter notebooks when you’re ready here.

You can tell if someone is using Python or iPython by looking at the prompt on the screen. Python always has a simple three arrow icon (>>>). iPython has a more colorful interface that looks like this:

In [1]: print('hello!')
hello!

Running Files: The other way to use Python – more common among software developers – is to write your python into a file, save it as something.py, then run that file by openning a command line interface and typing python something.py. This is analogous to openning Python, executing each line, then closing Python.

Don’t know what I mean by Command Line? Cool! You don’t need to know much about it to work with Python, but you should know enough you’re not scared of it. Take 10 minutes to run over to ST:Command Line for a quick overview.

Tutorials

Python 3 Essentials from Lynda.com

Pre-requisites: Some experience in another language like R or Java. If you’ve used for-loops and defined your own functions in another language, you’re probably fine.

The Good: A really incredibly clear instructor and carefully designed class. Some assumed familiarity with programming, but not a lot. Lynda.com also offers ability to speed-up playback for sections that feel familiar or straightforward. Topics are carefully organized and indexed if you need to jump around. Includes a “Quick Start” section.

The Bad: May not be free. Lynda.com tutorials are great because it’s a paid service, and sadly you sometimes get what you pay for. However, many universities (Yale, Stanford, etc.) have subscriptions, so check with your institution to see if you can get free access. You can also get a 10 day free trial, or pay $30 for a month. If you’re unsure, try the “Quick Start” section, which is public.

Recommended Sections: I would recommend proceeding as follows:

  • Section 1
  • Section 2 up to “Reusing code and data with a class”
  • If you installed Python using the setup recommended here, skip section 3.
  • Do Sections 4 - 8, 11, 13, 14, and the first two parts of Section 17

Optional Sections: Not immediately needed, but potentially quite useful:

  • Section 9

Other Notes: About a decade ago Python began a transition from version 2 to version 3 (for reasons that aren’t worth getting into, this was a somewhat controversial move, but it’s basically done at this point). This was published in part to help with that transition, so if you hear digressions about how Python 3 is different from Python 2, that’s why.

A Byte of Python

Pre-requisites: Zero assumed knowledge!

The Good: A really, really nice, free textbook for learning Python with no assumed knowledge of programming.

The Bad: Don’t have one! Very good.

Python in 10 Minutes

Pre-requisites: Knowing how to program in another language

The Good: If you’ve programmed a fair amount in another language, this is a great “get up and go” resource, but not ideal for beginners.

The Bad: If you haven’t programmed a lot in another language, not great. Also not great for mastery.

Python for Data Science

Pre-requisites: None, it appears!

The Good: A nice, focused, “let’s get going” text-based tutorial for social scientists. Doesn’t waste much time on things like classes, which I really appreciate!

The Bad: It talks a lot about differences between Python 2 and Python 3. As noted elsewhere on this site, you should really only work in Python 3. This is Python 3 focused, but you’ll have to wade through some junk about Python 2.

Python for You and Me

Pre-requisites: If you’ve done any programming in another language, you should be set. Maybe just a little too much assumed knowledge for someone who has never programmed, but I could be wrong on that (I’m very intolerant of assumed knowledge in teaching...).

The Good: If you don’t like video tutorials, this is a great choice. Clearly written, moves slowly and incrementally.

The Bad: No explicit exercises to work through.

Recommended Sections: I would recommend proceeding as follows:

  • Everything up to but not including “File Handling”
  • Modules

Optional Sections: Not crucial, but potentially quite helpful:

  • PEP8 Guidelines

Other Notes:

Automate the Boring Stuff

Pre-requisites: None! Though the name is a little weird, it seems like a great resource for social scientists.

The Good: Seems like a great introduction with essentially no assumed knowledge! The holy grail for absolute beginners. Also includes lectures (links to youtube at top of each section) for those who like it.

The Bad: The narrative voice is fun but a little verbose (kinda like this site), so it could feel a little slow for people with more background.

Recommended Section:

  • Chapters 0-6

Optional Sections: Not crucial, but potentially quite helpful:

  • Chapter 7, Chapter 10

A Note on Omitted Tutorials

Some users will note that I have left several relatively popular Python tutorials off this list. In most cases, this is due to the fact that I made an executive editorial decision early on in writing this website to make it “Python 3 Only”. “Uh, what?” you say?

About a decade ago, Python version 3 was released. Python 3 changed several problems that existed in Python 2, but as a result, code written in Python 2 would no longer work. For a while, both Python 2 and Python 3 were supported side by side, and the world has taken a while to transition to Python 3.

Many popular tutorials (like Python the Hard Way) are written in Python 2. As recently as a few years ago, this made sense because many important libraries weren’t yet available in Python 3. Moreover, if you plan to be a software developer (the target of many tutorials), it’s still absolutely necessarily you are familiar with both Python 2 and Python 3 since old code you will encounter at a company may still be in Python 2.

But for social scientists, I think it just makes sense to start off with Python 3. Almost all libraries (especially in data-science) are updated, and Guido van Rossum (Python’s “Benevolent Dictator for Life”) has made it very clear there will never be another Python 2 release. With that in mind, I’ve been avoiding any tutorial written in Python 2. I think that asking social scientist to learn a programming language is hard enough; also asking them to learn the language AND understand all the small differences between 2 and 3 was just a pointless invitation for confusion.