Python scripts for the analysis of supreme audit institution decisions, using latent Dirichlet allocation (LDA)

The Python scripts are divided into three parts:
- apply_method.ipynb uses developed functions to run LDA and analyse decisions in a Jupyter Notebook. These functions are found in the sai_data.py script.
- sai_data.py - contains functions that can be adapted to analyse decisions from different audit institutions as well as different terms.
- sai_collect.py script performs web crawling to retrieve the data from the Brazilian Supreme Audit Institution.
File requirements.txt contains the necessary packages to run the scripts.