Online Program

Return to main conference page

All Times ET

Friday, June 4
Practice and Applications
Data Science Shaping Innovative Applications
Fri, Jun 4, 11:25 AM - 1:00 PM
TBD
 

Improving Election Predictions: A Statistical Investigation of Crucial Swing State Behavior in the Past (309698)

Presentation

*Mason Chen, Stanford OHS 
Saloni Patel, Stanford OHS 

Keywords: Presidential Election, Swing States, Hierarchical Clustering, COVID-19

Although President Biden has been declared president-elect of the 2020 US presidential election, his road to victory was not a straightforward one, influenced by unexpected results from several swing states. This election was unique in many respects, but it was definitely not the first one to be impacted by results from crucial swing states. The main objective of this project is to analyze the impact of government decisions made from the Trump administration and other important events on election outcomes. First, the hierarchical clustering tool was used to group the 15 swing states based on the 2012, 2016, and 2020 election results, and the relationships between and within each cluster was further studied and attributed with events that may have factored into the cluster behavior. Next, a swing state index was devised to study the swing behavior of each cluster, in order to analyze how current situations may have factored into the election results of several states. The study of similar events and government decisions can be applied to future elections to better predict the outcome of important swing states.