This repo contains the code and data needed to reproduce the results in Turetsky & Riddle (2018).

To be fully reproducible, one also needs to download google's pre-trained word2vec word embeddings. They are located here. If interested in an end-to-end run, one should run the analysis files in the following order:

• edge_processing.py
• clean_edgelist.R
• compute_sentiment.py
• dict_methods.py
• spps_analysis.R

One should double check all file paths to ensure they are consistent with the directory structure on the users's machine.