Polypilot product mascot

Introducing PolyPilot:

Our AI-Powered Mentorship Program

Start your trial today

Learn More
Go to Polygence Scholars page
Ziang Huang's cover illustration
Polygence Scholar2024
Ziang Huang's profile

Ziang Huang

Shenzhen College of International EducationClass of 2023Shenzhen, Guangdong / 广东

About

Projects

  • "Who controls U.S. politics? An analysis of major political endorsements in U.S. midterm elections" with mentor Grayson (Working project)

Ziang's Symposium Presentation

Project Portfolio

Who controls U.S. politics? An analysis of major political endorsements in U.S. midterm elections

Started July 25, 2022

Abstract or project description

Former president Donald Trump endorsed over 200 candidates during the 2022 election cycle; in total, he has endorsed 551 candidates since he took office. Facilitated by the convenience of social media, other major political figures, including Bernie Sanders and Barack Obama, have endorsed similar numbers of candidates. Though such endorsements ostensibly advance a certain political cause and support a candidate’s viability, it is uncertain what effects they actually have on the political landscape of the United States today. In this project, we want to determine how much of a measurable effect these endorsements have on midterm elections in the United States and the political makeup of Congress as a whole. In order to quantify the effect of an endorsement on an election, we hope to perform a regression, with several fundamental factors, including endorsements, state and district demographics (race, income and age distribution), and political leanings by state. Data on these factors can be obtained from several sources, including Nate Silver's 538 sites, Ballotpedia, and the US Census Bureau. Once we obtain our data, we will perform a statistical analysis and construct a predictive model for the election outcome based on the fundamental variables.

We can use multiple statistical methods to make predictions based on our data, including linear regression, multiple regression, and random forests. We will also assess the robustness and predictive power of our model through “leave-one-out” cross-validation methods and applying election results of previously unseen candidates endorsed by major political figures. Through evaluating the endorsement impact of major political figures in the US, we hope to ascertain whether or not former presidents such as Trump or Obama still retain lingering influence over their party, as well as the “kingmakers” of Congressional elections today. Finally, we will apply our model to the ongoing 2022 midterm elections (featuring many candidates for which Biden, Sanders, Trump etc. have already staked their claim) and attempt to forecast each endorsed candidate’s victory margin through fundamental factors, state political leanings, and endorsement impacts.