Josh Cutler - Computational Frameworks for Political and Social Research with Python

This book introduces concepts from Computer Science and Software Engineering using the Python programming language. The goal is to give readers a strong working knowledge of the skills necessary to use programming in their social science research. We also survey more advanced topics to give readers familiarity with computational modeling and data analysis to serve as a starting point for further learning.

In recent years, knowledge of one or more programming languages has become a valuable part of the social science toolkit. One example is R, which is widely used in both statistics and the social sciences.As social science researchers wish to take on more data-intensive tasks (such as web scraping, machine learning, and image processing), we claim that incorporating Python into their workflow will become an increasingly valuable skill.

Although both R and Python can in theory be used for any programming task, the availability of well-documented and widely supported libraries makes some tasks easier in one ecosystem than another. In Chap.. Readers interested in pursuing other advanced technologies beyond the scope of this text, such as TensorFlow for machine learning and Spark for large-scale distributed computing, will also benefit from a knowledge of Python.

We do not argue in this text that social scientists should not learn R. Rather, we claim that once a researcher is comfortable using R, Python is a compelling choice for the next programming language to learn. It is also a more natural choice for illustrating fundamental concepts from computer science and computational social science. The ability to read Python code will also make work in these and other disciplines outside social science more accessible. Throughout this text we will make references to certain syntactic differences between the two languages to help readers familiar with R translate their knowledge to Python.

This book is recommended for people with some programming skill but no formal computer science education. We assume that readers are familiar with basic syntax and programming concepts (e.g., looping, conditionals, functions). We will not assume that readers are familiar with bigOcomplexity, data structures, databases, or any other specialized computer science topics.

Familiarity with Python and its concepts (object-oriented programming, OOP) will allow you to hit the ground running. Chapterintroduces these topics and provides references for further reading.

This book grew out of a course in the Political Science Department at Duke University. That course was designed to introduce graduate and doctoral students to some of the computational tools that they could use in their research. Initial versions of this course were designed to help social scientists create new and interesting datasets by leveraging the internet and previously ignored unstructured data sources.

This book can be used in either of two ways: as a self-teaching tool or as the basis for an advanced undergraduate- or graduate-level course.

To get the most out of the book for self-teaching purposes, we recommend starting with the introduction to Python in Chap.), even if you know conceptually how to perform the same task in a language other than Python. After completing the first section, readers can focus on chapters of interest in the latter section since these chapters do not build upon one another.

The second way to use this book, and the one that we focused on in writing it, is as the textbook for a one-semester (or one-quarter) course. Students in the course at Duke University were primarily graduate and doctoral students, though some upper-level undergraduates also participated. The modal programming background was familiarity with the R language, and some students also understood the basics of a markup language (e.g., HTML) or another programming language (e.g., Perl).