Doing Computational Social Science with Python and R

This session will familiarize students with the fundamentals of Python while simultaneously learning preprocessing operations for text data. First, there will be a brief introduction describing what text-as-data and NLP are and why they are used in the social sciences. Then, using a toy text dataset, students will learn about the fundamentals of Python through the following topics. Setup: pip, writing code in a text editor, running python code from the terminal, virtual environment basics (venv), importing modules. Variables and operations: strings, integers, floats, Booleans. Basic data structures: lists and dictionaries. Iteration: for loops, while loops, flow control. Reading and writing files. Functions and classes. These topics will be introduced through practical exercises on tokenizing text, removing stopwords, splitting text into chunks, integrating these tasks into functions and classes.

Any text editor for coding (e.g. Notepad++, Sublime Text, CotEditor, VS Code); pip; Python.

Before the workshop:

1. Download the latest stable version of Python here. Full instructions for installing Python for your particular operating system can be found here (see section 4 for Python on Windows, section 5 for Python on MacOS).
2. Ensure you have a text editor installed. Your computer should have a default one installed like Notepad or TextEdit – you can just use that. Alternatively, you can try other options, such as Notepad++ for Windows or CotEditor for Mac OS.
3. Once Python is installed, download and install pip. This can be done from the command line (command prompt in Windows, terminal in MacOS and Linux). Open your command line and run the following commands to install pip depending on your operating system:
   • Windows: py -m ensurepip --upgrade
   • Mac OS or Linux: python -m ensurepip --upgrade (if you get a “command not found” error, type python3 instead of python in this code)

In this session we will explain the intuition behind word embeddings through practical exercises. To grasp the intuition, students will first do exercises using pre-trained static embeddings (word2vec, fastext etc.) to see how distances between embeddings correspond to differences in the meanings of words. This will also get students used to numpy arrays as a type of data structure. Then, students will do exercises to practice training static word embeddings. In the second half of the session students will be introduced to contextual embeddings with BERT and will conduct exercises on (1) using contextual embeddings to perform topic modelling and (2) visualizing and describing the results of topic modelling.

numpy, gensim, transformers, bertopic, sentence transformers, matplotlib.

Before the workshop, make sure you have completed the setup for Session 1 and that you have Python and pip installed.

Then set up a python virtual environment with the following instructions. Don’t worry too much about what a virtual environment is — if you were unable to attend the first session, I will explain that in other sessions too:

1. Choose or create a folder where you want your virtual environment to be. For example, you might choose or create the folder C:\Users\YourUsername\Documents\Projects\my_python_project
2. Open your command line (command prompt in Windows, terminal in MacOS and Linux) and use the cd command to navigate to the folder you selected/created in point (1) within the command line, for example: cd C:\Users\YourUsername\Documents\Projects\my_python_project
3. Use the following command to create a virtual environment named env1: python -m venv env1 (if you get a “command not found” error, type python3 instead of python in this code)
4. You now have a virtual environment called env1 in the folder you selected/created in point (1). Every time you need to activate env1, you’ll need to make sure you have navigated to the folder containing it using cd, and then you’ll need to enter the following command:
   • (For Windows): env1\Scripts\activate
   • (For MacOS/Linux): source env1/bin/activate
5. Make sure you have activated env1. Then run the following commands:
   • pip install gensim
   • pip install numpy
   • pip install bertopic

The first half of this session will teach students the fundamentals of API usage with exercises based on Youtube’s API. Exercises will be designed to get students familiar with the steps common to API usage, authorization, each API’s Python interface, and writing data in dictionary form to JSON files. The second half will be focused on web scraping. Here, students will learn how to make Python requests, how to inspect webpage elements that contain useful data one might want to collect using their browsers’ inspection tool, and how to select elements using BeautifulSoup. Students will also be taught some of the practical considerations of web scraping, such as making your code pause between each request to avoid server overload, understanding that scraping some websites might not be permitted, etc.

requests, json, beautifulsoup, os.

Before the workshop, make sure you have completed the setup for Session 1 and that you have Python and pip installed.

Then set up a python virtual environment with the following instructions. Don’t worry too much about what a virtual environment is — if you were unable to attend the first session, I will explain that in other sessions too:

1. Choose or create a folder where you want your virtual environment to be. For example, you might choose or create the folder C:\Users\YourUsername\Documents\Projects\my_python_project
2. Open your command line (command prompt in Windows, terminal in MacOS and Linux) and use the cd command to navigate to the folder you selected/created in point (1) within the command line, for example: cd C:\Users\YourUsername\Documents\Projects\my_python_project
3. Use the following command to create a virtual environment named env1: python -m venv env1
4. You now have a virtual environment called env1 in the folder you selected/created in point (1). Every time you need to activate env1, you’ll need to make sure you have navigated to the folder containing it using cd, and then you’ll need to enter the following command:
   • (For Windows): env1\Scripts\activate
   • (For MacOS/Linux): source env1/bin/activate
5. Make sure you have activated env1. Then run the following commands:
   • pip install --upgrade google-api-python-client
   • pip install --upgrade google-auth-oauthlib google-auth-httplib2
   • pip install beautifulsoup4
6. Follow instructions here (under “Before you start”).  This is needed to use the Youtube API, which is what we will be practising with in the session.

1. Download the latest version of R here (select a location near you)
   • Follow instructions to install R in your computer
   • If you already downloaded R some time ago, please update it to the latest version
2. Download the latest version of RStudio (free version) here
   • Follow instructions to install RStudio in your computer
   • If you already downloaded RStudio some time ago, please update it to the latest version
3. Install the following R packages: ergm, sna, network, igraph

1. Download the latest version of R here (select a location near you)
   • Follow instructions to install R in your computer
   • If you already downloaded R some time ago, please update it to the latest version
2. Download the latest version of RStudio (free version) here
   • Follow instructions to install RStudio in your computer
   • If you already downloaded RStudio some time ago, please update it to the latest version
3. Install the following R packages: ergm, sna, network, igraph

This session has two primary objectives. In the first half, students will explore key features of geospatial data, including handling polygon data, importing and exporting GeoJSON data, and integrating geospatial data with Pandas DataFrames. Additionally, the session will cover essential tools for Python programming, such as Conda (Anaconda), JupyterLab, Jupyter Notebook, and creating virtual environments and managing package versions with YAML files. The second half of the session focuses on applying practical skills using a dataset of Airbnb reviews from Milan. Here, students will learn how to process and visualize geospatial data in Python. They will also engage in hands-on exercises for topic modeling of accommodation review comments, along with the visualization of these results on an interactive map.

Install Anaconda. Additional materials will be provided prior to the workshop.

This session assumes that participants have completed the previous sessions, especially Session 1 and Session 2. If you have not, please ensure you are familiar with the following:

1. What a virtual environment is
2. How to install Python libraries using pip

Setup instructions for this session:

1. Download the latest stable version of Anaconda here.
2. Visit Inside Airbnb and find “Milan.” Download all files into your working directory.
3. Go here and download css-python-geo.yml.
4. Open your command line (for Windows users, please use Anaconda Prompt) and run the following commands to create a pre-defined virtual environment using the downloaded file:
   • conda env create -f css-python-geo.yml
   • conda activate css-python-geo

This session assumes that participants have completed the previous sessions, especially Session 1 and Session 2. If you have not, please ensure you are familiar with the following:

1. What a virtual environment is
2. How to install Python libraries using pip

Setup instructions for this session:

1. To install the mesa library, open your command line (for Windows users, use Anaconda Prompt) and run the following code:
   • pip install -U mesa