Authors:

Joseph Michael Cromp

Date Created:

Not specified

Course Title:

Professor:

Not specified

About Paper:

Forecasting elections is an exciting, yet complicated, endeavor that is often scrutinized by the public. Additionally, the methods of popular forecasters are not always fully transparent, making it difficult to understand their methods. Our research aims to accurately forecast the 2024 U.S. presidential, senatorial, and gubernatorial elections and share our results with the public. We use a Susceptible-Infected-Susceptible model to track interactions between Democratic, Republican, and undecided voters. Using polling data, we produce ten thousand possible election-night scenarios that make up our forecasts. In past election cycles, this method performed as well as major forecasters such as Sabato's Crystal Ball and FiveThirtyEight. To improve our future forecasts, we are also analyzing information about the polls that we use such as their sample size and associated polling organization. Using the results of our analysis, we will tweak our methods to improve our forecasts' accuracy. The final focus of our project is to visualize uncertainty by posting our forecasts and interactive tools on our website (https://c-r-u-d.gitlab.io/2024/) in the months leading up to Election Day 2024. Keywords: Mathematical Modeling; Forecasting Elections

Source:

Purdue University / 2024

Topics:

No topics listed

Co-authors:

Joseph Michael Cromp