Can Presidential 'Executive Order' Volume Be Predicted?

Building a pipeline and predictive model of executive order counts based on economic and political indicators.

Overview

This is mostly a data engineering project: scraping data sources from a RESTful API, a regularly updated public-facing bank of CSV files, and a bunch of reasonably well-organized HTML tables. This data is then used to build a SARIMAX model, with hyperparameter tuning performed using mlflow. All of this work was performed on the Google Cloud Platform, because it's what I use at work right now, and this gave me the opportunity to play around with a bunch of stuff before I put in a big access request to our IAM team.

What this repository is emphatically not is an instruction manual. I have attached a free-to-use license here, and you're welcome to go to town with it, but this is a personal project (which I used for a capstone presentation during my Master's Degree.)

Data Architecture

Data Sources

Executive order data is retrieved from the Federal Register¹.

Economic indicator data is retreived from the U.S. Census Bureau².

Presidential job approval data is retreived from UC Santa Barbara³.

Acknowledgments

This work was heavily influenced by a letter⁴ published in the American Political Science Review in 2019 and, in the future, I hope to expand my work to include the indicators and methods outlined by Drs. Dino P. Christensen and Douglas L. Kriner.

Any success implementing an mlflow server on PostGRESQL on Google Cloud Platform is a result of the outstanding work provided here by DLabs. It is a little out of date, but links to a fairly current git repo with clear instructions. Connecting to the PostGRES instance is much easier this way using psycopg2 than using the pg8000 unix sock method.

References

Footnotes

1. U.S. National Archives (n.d.). Executive Orders. Federal Register. Retrieved January 21, 2025, from https://www.federalregister.gov/presidential-documents/executive-orders ↩

2. U.S. Census. (2025, January 21). Census Bureau Economic Briefing Room. Retrieved January 21, 2025, from https://www.census.gov/economic-indicators/ ↩

3. University of California Santa Barbara (2024, September 23). Presidential Job Approval--All Data. The American Presidency Project. Retrieved January 21, 2025, from https://www.presidency.ucsb.edu/statistics/data/presidential-job-approval-all-data ↩

4. CHRISTENSON, D. P., & KRINER, D. L. (2019). Does Public Opinion Constrain Presidential Unilateralism? American Political Science Review, 113(4), 1071–1077. doi:10.1017/S0003055419000327 ↩