Dampening Natural Disasters’ Disruptive Effects on Firms and Labor Markets

This paper seeks to understand the effects of climate shocks on firms and labor markets in Brazil.

Working with Daniela Scur, I created a data pipeline from scratch that did the following:

1. Web-scrape differently-structured ~5500 Brazilian municipality websites written in Portuguese, and download all the PDFs available (approximately 25 million PDFs).
2. The pipeline not only adapted to different website structures, but also evaded CAPTCHA issues using various techniques.
3. Using PDFPlumber, I converted information from these differently structured PDFs into usefully structured Excel files.
4. Additionally, I did geospatial data visualization in Python & Tableau to create interactive graphics for this project.
5. We are now conducting preliminary data analyses to understand the functions of municipal governments in Brazil.
6. I supervise the data collection for the project - which includes implementing machine learning and clustering algorithms, managing and tracking big data, natural language processing, and the use of LLMs to speed up data processing and reduce costs.

We are currently working the data collection and processing phase of this project, while we refine our research plan. Meanwhile, I assisted with the creation of a discrete choice experiment for this research as well.