Causal Inference Project

• NOTE: I used the pyenv, and pyenv-virtualenv to set up my python environment.
• For R I set it up using renv
  • I used R version 4.2.2

Setup

To run this first set up the project as follows:

git clone https://github.com/peterjbachman/causal_project.git
cd causal_project
pyenv virtualenv 3.11.1 causal_project
pyenv local causal_project
pip install -r requirements.txt

Set up R for this project by then running the following in the same terminal:

R

You will need to create a python script in this directory named secrets_cl.py formatted as follows:

api_key = {"Authorization": "Token <insert CAP API token here>"}

Replication

• 01_pull_data.py
  • Pulls the opinion text using the CAP API
• 02_clean_text.py
  • Cleans up the text and cleans the opinion author section
• 03_tidy_text.R
  • Cleans up the data further by trimming variables in the dataset and add a few variables
• 04_split_cases.R
  • Splits the data into a training and test dataset to learn about the topics in this data.
• 05_train_stm.R
  • Train Structural Topic Model on the training data, perform diagnostics and explore topics
• 06_pull_validate.R
  • Randomly sample 50 paragraphs, realize some stuff, only sample 49 documents and then hand code and validate the topics they were assigned to.
• 07_final_results.R
  • Run Structural Topic Model on testing data, Create Figure 1 on the poster
• 08_map.R
  • Create Figure 2 on the poster